Terminal-Bench

gpt-5.6-sol

OpenAI

closed

Source carrier

91.9%

Claude Opus 5

claude-opus-5

Anthropic

closed

Score input

89.1%

Grok 4.6

grok-4.6

xAI

closed

Score input

88.4%

kimi-k3

Moonshot AI

open

Score input

88.3%

GLM-5.3

glm-5.3

Z.AI

announced-open

Score input

88.2%

gpt-5.6-terra

OpenAI

closed

Score input

88.0%

claude-fable-5

Anthropic

closed

Public reference

88%

claude-mythos-5

Anthropic

closed

Source carrier

88%

DeepSeek-V4-Pro-0813

DeepSeek

open

Public reference

87.9%

#10

Qwen3.8 Max

qwen3.8-max

Alibaba Cloud

open

Score input

86.6%

#11

Qwen3.8-Flash-Next

Qwen/Qwen3.8-Flash-Next

Alibaba Cloud

open

Public reference

86.1%

#12

Gemini 3.7 Flash

gemini-3.7-flash

Google

closed

Score input

85.8%

#13

claude-opus-4-8

Anthropic

closed

Public reference

85%

#14

gpt-5.6-luna

OpenAI

closed

Source carrier

84.7%

#15

DeepSeek V4 Flash Vision Exp

gpt-5.5

OpenAI

closed

Public reference

84.3%

#16

GLM-5.3-Flash

glm-5.3-flash

Z.AI

open

Score input

84.3%

#17

deepseek-v4-flash-vision-exp

DeepSeek

—

Public reference

83.9%

#18

grok-4.5

xAI

closed

Source carrier

83.3%

#19

claude-opus-4-7

Anthropic

closed

Score input

83.1%

#20

muse-spark-1-2

Meta

announced-open

Score input

82.9%

#21

DeepSeek V4 Flash 0731

deepseek/deepseek-v4-flash-0731

DeepSeek

open

Public reference

82.7%

#22

Sakana Fugu-Ultra

sakana-fugu-ultra

Sakana AI

closed

Source carrier

82.1%

#23

SWE-1.7

swe-1-7

Cognition

closed

Source carrier

81.5%

#24

glm-5-2

Z.AI

open

Source carrier

81%

#25

claude-sonnet-5

Anthropic

closed

Score input

80.5%

#26

Sakana Fugu

sakana-fugu

Sakana AI

closed

Source carrier

80.2%

#27

muse-spark-1-1

Meta

closed

Score input

80%

#28

gpt-5.4

OpenAI

closed

Public reference

78.3%

#29

gemini-3.6-flash

Google

closed

Score input

78%

#30

Ornith-1.0-397B

ornith-1-0-397b

DeepReinforce AI

open

Source carrier

77.5%

#31

GPT-5.3-Codex

gpt-5.3-codex

OpenAI

closed

Source carrier

77.3%

#32

gemini-3.5-flash

Google

closed

Score input

76.2%

#33

qwen3.7-max

Alibaba Cloud

closed

Public reference

74.5%

#34

gemini-3.1-pro-preview

Google

closed

Public reference

73.8%

#35

Qwen3.8-27B

Qwen/Qwen3.8-27B

Alibaba Cloud

open

Score input

73%

#36

deepseek-v4-pro

DeepSeek

open

Public reference

72.1%

#37

LongCat-2.0

meituan-longcat/LongCat-2.0

Meituan

open

Score input

70.8%

#38

qwen3.7-plus

Alibaba Cloud

closed

Source carrier

70.3%

#39

Laguna S 2.1

laguna-s-2-1

Poolside

open

Source carrier

70.2%

#40

Composer 2.5

composer-2-5

Cursor

closed

Source carrier

69.3%

#41

mimo-v2-5-pro

Xiaomi

open

Source carrier

68.4%

#42

Kimi K2.7 Code

kimi-k2.7-code

Moonshot AI

open

Public reference

67.4%

#43

kimi-k2.6

Moonshot AI

open

Source carrier

66.7%

#44

MiniMax-M3

MiniMax

open

Source carrier

66%

#45

mimo-v2.5

Xiaomi

open

Source carrier

65.8%

#46

claude-opus-4-6

Anthropic

closed

Source carrier

65.4%

#47

Qwen 3.6 Max (preview)

qwen3-6-max-preview

Alibaba Cloud

closed

Source carrier

65.4%

#48

Qwen3.6 Max

qwen3.6-max

Alibaba Cloud

closed

Score input

65.4%

#49

Ornith-1.0-35B

ornith-1-0-35b

DeepReinforce AI

open

Source carrier

64.2%

#50

Inkling

inkling

Thinking Machines Lab

open

Source carrier

63.8%

#51

glm-5-1

Z.AI

open

Source carrier

63.5%

#52

deepseek-v4-flash

DeepSeek

open

Public reference

61.8%

#53

Composer 2

composer-2

Cursor

closed

Source carrier

61.7%

#54

Qwen3.6 Plus

qwen3-6-plus

Alibaba Cloud

closed

Source carrier

61.6%

#55

qwen3.6-27b

Alibaba Cloud

open

Public reference

60.7%

#56

gpt-5.4-mini

OpenAI

closed

Source carrier

60%

#57

Step 3.7 Flash

step-3-7-flash

StepFun

open

Source carrier

59.5%

#58

Claude Opus 4.5

claude-opus-4-5

Anthropic

closed

Source carrier

59.3%

#59

Claude Sonnet 4.6

claude-sonnet-4-6

Anthropic

closed

Source carrier

59.1%

#60

Muse Spark

muse-spark

Meta

closed

Source carrier

59%

#61

MiniMax M2.7

minimax-m2-7

MiniMax

open

Source carrier

57%

#62

nemotron-3-ultra-550b

NVIDIA

open

Source carrier

56.4%

#63

GLM-5

glm-5

Z.AI

open

Source carrier

56.2%

#64

claude-sonnet-4-5

Anthropic

closed

Score input

55.8%

#65

Hy3

hy3

Tencent

open

Score input

54.4%

#66

Hy3 Preview

hy3-preview

Tencent

open

Source carrier

54.4%

#67

gemini-3.5-flash-lite

Google

closed

Public reference

54%

#68

Qwen/Qwen3.5-397B-A17B

Alibaba Cloud

open

Source carrier

52.5%

#69

Muse Glimmer 30B

meta-models/Muse-Glimmer-30B

Meta

open

Score input

51.7%

#70

Qwen/Qwen3.6-35B-A3B

Alibaba Cloud

open

Source carrier

51.5%

#71

kimi-k2.5

Moonshot AI

open

Source carrier

50.8%

#72

Mistral Medium 3.5 128B

mistralai/Mistral-Medium-3.5-128B

Mistral AI

open

Score input

50.6%

#73

Mistral Medium 3.5

mistral-medium-3-5

Mistral AI

open

Public reference

50.6%

#74

GLM-4.6

glm-4.6

Z.AI

open

Score input

49.4%

#75

Qwen/Qwen3.5-122B-A10B

Alibaba Cloud

open

Source carrier

49.4%

#76

Qwen3.5 122B

qwen3.5-122b-a10b

Alibaba Cloud

open

Score input

47.6%

#77

Grok 4.20

grok-4.20

xAI

closed

Source carrier

47.1%

#78

DeepSeek V3.2

deepseek-v3.2

DeepSeek

open

Score input

46.8%

#79

GPT-5.4 nano

gpt-5.4-nano

OpenAI

closed

Source carrier

46.3%

#80

MAI-Thinking-1

mai-thinking-1

Microsoft

closed

Source carrier

46%

#81

Laguna M.1

laguna-m-1

Poolside

closed

Source carrier

45.8%

#82

NVIDIA Nemotron 3.5 Lightning 30B-A3B

glm-4.7

Z.AI

open

Public reference

45.3%

#83

DeepSeek V3.1 Terminus

deepseek-v3.1-terminus

DeepSeek

open

Score input

44.9%

#84

Ornith-1.0-9B

ornith-1-0-9b

DeepReinforce AI

open

Source carrier

43.1%

#85

Qwen3.5 27B

qwen3.5-27b

Alibaba Cloud

open

Source carrier

41.6%

#86

Qwen3.5-35B-A3B

qwen3-5-35b-a3b

Alibaba Cloud

open

Source carrier

40.5%

#87

Grok 4.3

grok-4.3

xAI

closed

Public reference

39.7%

#88

Gemma 4 26B A4B

google/gemma-4-26b-a4b-it

Google

open

Score input

39.0%

#89

Gemma 4 26B

gemma-4-26b-a4b

Google

open

Score input

39.0%

#90

Nemotron 3 Super

nemotron-3-super-120b

NVIDIA

open

Score input

38.6%

#91

Claude Sonnet 4

claude-sonnet-4

Anthropic

closed

Score input

36.3%

#92

Laguna XS.2

laguna-xs-2

Poolside

open

Source carrier

35.7%

#93

GPT-5

gpt-5

OpenAI

closed

Public reference

35.2%

#94

Gemini 3.1 Flash-Lite

gemini-3.1-flash-lite

Google

closed

Public reference

31%

#95

Nova 2 Pro

nova-2-pro-preview

Amazon

closed

Score input

29.6%

#96

Gemini 2.5 Pro

gemini-2.5-pro

Google

closed

Public reference

28.5%

#97

nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16

NVIDIA

open

Score input

24.6%

#98

Command A+

command-a-plus

Cohere

open

Public reference

22.8%

#99

Mistral Small 4

mistral-small-2603

Mistral AI

open

Score input

21.0%

#100

Trinity-Large-Thinking

trinity-large-thinking

Arcee AI

open

Score input

20.6%

#101

Mistral Large 3

mistral-large-2512

Mistral AI

open

Public reference

12.0%

#102

GPT-4.1 mini

gpt-4-1-mini

OpenAI

closed

Score input

10.1%

#103

Llama 4 Maverick

llama-4-maverick

Meta

open

Score input

7.9%

#104

Nemotron 3 Nano 30B

nemotron-3-nano-30b

NVIDIA

open

Score input

6.7%

#105

Gemma 3 27B

gemma-3-27b

Google

open

Score input

4.5%

#106

GPT-4.1 nano

gpt-4-1-nano

OpenAI

closed

Score input

3.7%

#107

Llama 4 Scout

llama-4-scout

Meta

open

Score input

3.7%

Every distinct dated source-published row for Terminal-Bench, including exact model configuration labels
Exact published model / configuration	Canonical product	Score	Version / system	Evidence	Source / observed
Qwen3.8-Flash-Next (Artificial Analysis independent run) qwen-3-8-flash-next-aa-unspecified	Qwen3.8-Flash-Next	86.1%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Public reference · not scored independently-verified	Qwen3.8-Flash-Next individual evaluations ↗Observed 2026-08-29 · checked 2026-08-29
Claude Opus 4.7 (Adaptive Reasoning, Max Effort) (Artificial Analysis independent run) claude-opus-4-7-max	Claude Opus 4.7	83.1%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Claude Opus 4.7 (Adaptive Reasoning, Max Effort) individual evaluations ↗Observed 2026-08-29 · checked 2026-08-29
GLM-5.2 (max) (Artificial Analysis independent run) glm-5-2-max	GLM-5.2	77.9%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	GLM-5.2 (max) individual evaluations ↗Observed 2026-08-29 · checked 2026-08-29
Qwen3.5 397B A17B (Reasoning) (Artificial Analysis completed independent run) qwen3-5-397b-thinking	Qwen3.5 397B A17B	51.3%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Qwen3.5 397B A17B (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
Mistral Medium 3.5 (Artificial Analysis completed independent run) mistral-medium-3-5-128b-aa-reasoning-default	Mistral Medium 3.5 128B	50.6%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Mistral Medium 3.5 current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
LongCat 2.0 (Artificial Analysis independent run) longcat-2-0-default	LongCat-2.0	50.2%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	LongCat 2.0 individual evaluations ↗Observed 2026-08-29 · checked 2026-08-29
Qwen3.5 122B A10B (Reasoning) (Artificial Analysis completed independent run) qwen3-5-122b-a10b-aa-reasoning-default	Qwen3.5-122B-A10B	47.6%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Qwen3.5 122B A10B (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
Kimi K2.5 (Reasoning) (Artificial Analysis completed independent run) kimi-k2-5-thinking	Kimi K2.5	45.7%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Kimi K2.5 (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
Qwen3.6 35B A3B (Reasoning) (Artificial Analysis completed independent run) qwen3-6-35b-a3b-aa-reasoning-default	Qwen3.6-35B-A3B	44.9%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Qwen3.6 35B A3B (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
Gemma 4 26B A4B (Reasoning) (Artificial Analysis completed independent run) gemma-4-26b-a4b-aa-reasoning-default	Gemma 4 26B A4B	39.0%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Gemma 4 26B A4B (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
Mistral Small 4 (Reasoning) (Artificial Analysis completed independent run) mistral-small-4-reasoning	Mistral Small 4	21.0%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Mistral Small 4 (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
Trinity Large Thinking (Artificial Analysis completed independent run) trinity-large-thinking-aa-reasoning-default	Trinity-Large-Thinking	20.6%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Trinity Large Thinking current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
GPT-4.1 mini (Artificial Analysis completed independent run) gpt-4-1-mini-epoch-gpt-4-1-mini-2025-04-14	GPT-4.1 mini	10.1%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	GPT-4.1 mini current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
Llama 4 Maverick (Artificial Analysis completed independent run) llama-4-maverick-aa-non-reasoning-default	Llama 4 Maverick	7.9%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Llama 4 Maverick current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
NVIDIA Nemotron 3 Nano 30B A3B (Reasoning) (Artificial Analysis completed independent run) nemotron-3-nano-30b-aa-reasoning-default	Nemotron 3 Nano 30B	6.7%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	NVIDIA Nemotron 3 Nano 30B A3B (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
Gemma 3 27B Instruct (Artificial Analysis completed independent run) gemma-3-27b-aa-non-reasoning-default	Gemma 3 27B	4.5%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Gemma 3 27B Instruct current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
GPT-4.1 nano (Artificial Analysis completed independent run) gpt-4-1-nano-epoch-gpt-4-1-nano-2025-04-14	GPT-4.1 nano	3.7%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	GPT-4.1 nano current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
Llama 4 Scout (Artificial Analysis completed independent run) llama-4-scout-aa-non-reasoning-default	Llama 4 Scout	3.7%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Llama 4 Scout current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29
Claude Opus 5 (max; Artificial Analysis independent run) claude-opus-5-max	Claude Opus 5	89.1%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Claude Opus 5 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
Grok 4.6 (high; Artificial Analysis independent run) grok-4-6-high	Grok 4.6	88.4%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Grok 4.6 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
GPT-5.6 Sol (max; Artificial Analysis independent run) gpt-5-6-sol-max	GPT-5.6 Sol	88.0%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	GPT-5.6 Sol individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
GPT-5.6 Terra (max; Artificial Analysis independent run) gpt-5-6-terra-max	GPT-5.6 Terra	88.0%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	GPT-5.6 Terra individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
Kimi K3 (max; Artificial Analysis independent run) kimi-k3-max	Kimi K3	85.0%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Kimi K3 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
Claude Fable 5 (max; Artificial Analysis independent run) claude-fable-5-max	Claude Fable 5	84.6%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Claude Fable 5 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
Claude Opus 4.8 (max; Artificial Analysis independent run) claude-opus-4-8-max	Claude Opus 4.8	84.6%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Claude Opus 4.8 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
GLM-5.3-Flash (max; Artificial Analysis independent run) glm-5-3-flash-max	GLM-5.3-Flash	84.3%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	GLM-5.3-Flash individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
GPT-5.5 (xhigh; Artificial Analysis independent run) gpt-5-5-xhigh	GPT-5.5	84.3%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	GPT-5.5 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
GLM-5.3 (max; Artificial Analysis independent run) glm-5-3-max	GLM-5.3	83.9%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	GLM-5.3 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
Grok 4.5 (high; Artificial Analysis independent run) grok-4-5-aa-2-high	Grok 4.5	81.6%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Grok 4.5 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
GPT-5.6 Luna (max; Artificial Analysis independent run) gpt-5-6-luna-max	GPT-5.6 Luna	80.9%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	GPT-5.6 Luna individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
Claude Sonnet 5 (max; Artificial Analysis independent run) claude-sonnet-5-max	Claude Sonnet 5	80.5%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Claude Sonnet 5 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
Muse Spark 1.2 (xhigh; Artificial Analysis independent run) muse-spark-1-2-xhigh	Muse Spark 1.2	80.1%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Muse Spark 1.2 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
DeepSeek V4 Pro 0813 (max; Artificial Analysis independent run) deepseek-v4-pro-0813-max	DeepSeek V4 Pro 0813	78.7%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Public reference · not scored independently-verified	DeepSeek V4 Pro 0813 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
Gemini 3.7 Flash (medium; Artificial Analysis independent run) gemini-3-7-flash-medium	Gemini 3.7 Flash	78.3%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Gemini 3.7 Flash individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
GPT-5.4 (xhigh; Artificial Analysis independent run) gpt-5-4-xhigh	GPT-5.4	78.3%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	GPT-5.4 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
Muse Spark 1.1 (xhigh; Artificial Analysis independent run) muse-spark-1-1-xhigh	Muse Spark 1.1	77.9%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Muse Spark 1.1 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
Gemini 3.6 Flash (high; Artificial Analysis independent run) gemini-3-6-flash-high	Gemini 3.6 Flash	77.5%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Score input independently-verified	Gemini 3.6 Flash individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
DeepSeek V4 Flash Vision Exp (max; Artificial Analysis independent run) deepseek-v4-flash-vision-exp-max-harness	DeepSeek V4 Flash Vision Exp	74.2%	2.1Artificial Analysis Terminal-Bench v2.1 in e2b	Public reference · not scored independently-verified	DeepSeek V4 Flash Vision Exp individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27
DeepSeek Harness Minimal Mode; max effort; top_p=0.95; temperature=1.0 deepseek-v4-flash-vision-exp-max-harness	DeepSeek V4 Flash Vision Exp	83.9%	2.1DeepSeek Harness Minimal Mode	Public reference · not scored provider-reported	DeepSeek-V4-Flash-Vision-Exp release and provider evaluation ↗Observed 2026-08-21 · checked 2026-08-24
Claude Opus 4.6 Max as published by Qwen claude-opus-4-6-max	Claude Opus 4.6	78.2%	2.1Qwen3.8-27B official text table	Relative comparison · not standard direct provider-reported	Qwen3.8-27B official model card ↗Observed 2026-08-15 · checked 2026-08-15
Qwen3.8-27B (xhigh) qwen-3-8-27b-xhigh	Qwen3.8-27B	73%	2.1Qwen3.8-27B official text table	Score input provider-reported	Qwen3.8-27B official model card ↗Observed 2026-08-15 · checked 2026-08-15
Qwen3.7-Plus as published by Qwen qwen-3-7-plus-unspecified	Qwen3.7-Plus	64%	2.1Qwen3.8-27B official text table	Relative comparison · not standard direct provider-reported	Qwen3.8-27B official model card ↗Observed 2026-08-15 · checked 2026-08-15
Qwen3.6-27B as published by Qwen qwen3-6-27b-default	Qwen3.6 27B	63.4%	2.1Qwen3.8-27B official text table	Relative comparison · not standard direct provider-reported	Qwen3.8-27B official model card ↗Observed 2026-08-15 · checked 2026-08-15
GPT-5.6 Sol as published by Z.AI gpt-5-6-sol-max	GPT-5.6 Sol	88.8%	2.1Z.AI GLM-5.3 official comparison table	Relative comparison · not standard direct provider-reported	GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15
Kimi K3 as published by Z.AI kimi-k3-max	Kimi K3	88.3%	2.1Z.AI GLM-5.3 official comparison table	Relative comparison · not standard direct provider-reported	GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15
GLM-5.3 (max) glm-5-3-max	GLM-5.3	88.2%	2.1Z.AI GLM-5.3 official comparison table	Score input provider-reported	GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15
Claude Fable 5 w/ fallback as published by Z.AI claude-fable-5-max	Claude Fable 5	88%	2.1Z.AI GLM-5.3 official comparison table	Relative comparison · not standard direct provider-reported	GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15
DeepSeek V4 Pro 0813 as published by Z.AI deepseek-v4-pro-0813-max	DeepSeek V4 Pro 0813	87.9%	2.1Z.AI GLM-5.3 official comparison table	Relative comparison · not standard direct provider-reported	GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15
Qwen3.8 Max as published by Z.AI qwen-3-8-max-xhigh	Qwen3.8 Max	86.6%	2.1Z.AI GLM-5.3 official comparison table	Relative comparison · not standard direct provider-reported	GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15
Claude Opus 4.8 as published by Z.AI claude-opus-4-8-max	Claude Opus 4.8	85%	2.1Z.AI GLM-5.3 official comparison table	Relative comparison · not standard direct provider-reported	GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15
GLM-5.2 as published by Z.AI glm-5-2-max	GLM-5.2	81%	2.1Z.AI GLM-5.3 official comparison table	Relative comparison · not standard direct provider-reported	GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15
Kimi K3 kimi-k3-deepseek-0813-release-unspecified	Kimi K3	88.3%	2.1DeepSeek V4 Pro 0813 official GA comparison table	Public reference · not scored provider-reported	DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
Fable 5 (with fallback) claude-fable-5-deepseek-0813-release-unspecified	Claude Fable 5	88%	2.1DeepSeek V4 Pro 0813 official GA comparison table	Public reference · not scored provider-reported	DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
DeepSeek V4 Pro 0813 deepseek-v4-pro-0813-deepseek-0813-release-unspecified	DeepSeek V4 Pro 0813	87.9%	2.1DeepSeek V4 Pro 0813 official GA comparison table	Public reference · not scored provider-reported	DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
GPT-5.6 Terra gpt-5-6-terra-max	GPT-5.6 Terra	87.4%	2.1Terminal-Bench 2.1 default Terminus 2 agent harness	Public reference · not scored provider-reported	Google DeepMind permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
Gemini 3.7 Flash gemini-3-7-flash-medium	Gemini 3.7 Flash	85.8%	2.1Terminal-Bench 2.1 default Terminus 2 agent harness	Score input provider-reported	Google DeepMind permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
Claude Opus 4.8 claude-opus-4-8-deepseek-0813-release-unspecified	Claude Opus 4.8	85%	2.1DeepSeek V4 Pro 0813 official GA comparison table	Public reference · not scored provider-reported	DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
Muse Spark 1.2 muse-spark-1-2-xhigh	Muse Spark 1.2	82.9%	2.1Terminal-Bench 2.1 default Terminus 2 agent harness	Public reference · not scored provider-reported	Google DeepMind permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
DeepSeek V4 Flash 0731 deepseek-v4-flash-0731-deepseek-0813-release-unspecified	DeepSeek V4 Flash 0731	82.7%	2.1DeepSeek V4 Pro 0813 official GA comparison table	Source carrier · no additional scoring weight provider-reported	DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
GLM-5.2 glm-5-2-deepseek-0813-release-unspecified	GLM-5.2	81%	2.1DeepSeek V4 Pro 0813 official GA comparison table	Public reference · not scored provider-reported	DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
Claude Sonnet 5 claude-sonnet-5-max	Claude Sonnet 5	80.4%	2.1Terminal-Bench 2.1 default Terminus 2 agent harness	Public reference · not scored provider-reported	Google DeepMind permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
Gemini 3.6 Flash gemini-3-6-flash-high	Gemini 3.6 Flash	78%	2.1Terminal-Bench 2.1 default Terminus 2 agent harness	Public reference · not scored provider-reported	Google DeepMind permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
DeepSeek V4 Pro Preview deepseek-v4-pro-deepseek-0813-release-unspecified	DeepSeek V4 Pro	72.1%	2.1DeepSeek V4 Pro 0813 official GA comparison table	Public reference · not scored provider-reported	DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
DeepSeek V4 Flash Preview deepseek-v4-flash-deepseek-0813-release-unspecified	DeepSeek V4 Flash	61.8%	2.1DeepSeek V4 Pro 0813 official GA comparison table	Public reference · not scored provider-reported	DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16; NVIDIA release evaluation; temperature=1.0; top_p=0.95 nemotron-3-5-lightning-30b-a3b-default	NVIDIA Nemotron 3.5 Lightning 30B-A3B	24.6%	2.1NVIDIA NeMo Gym / NeMo Evaluator SDK consistent release harness	Score input provider-reported	NVIDIA Nemotron 3.5 Lightning 30B-A3B BF16 model card ↗Observed 2026-08-11 · checked 2026-08-12
Muse Glimmer-30B; high reasoning; temperature=1.0; top_p=0.95; top_k=64 muse-glimmer-30b-high	Muse Glimmer 30B	51.7%	2.1Artificial Analysis Terminal-Bench 2.1 evaluation in an E2B sandbox	Public reference · not scored provider-reported	Muse Glimmer Evaluation Methodology ↗Observed 2026-08-10 · checked 2026-08-10
Claude Opus 5 (reasoning configuration not stated in chart) claude-opus-5-meta-muse-12-release-unspecified	Claude Opus 5	86.7%	2.1Meta release chart for Terminal-Bench 2.1; the chart labels each agent scaffold separately	Relative comparison · not standard direct provider-reported	Meta Superintelligence Labs permanent refresh source ↗Observed 2026-08-05 · checked 2026-08-15
Muse Spark 1.2 (xhigh) muse-spark-1-2-xhigh	Muse Spark 1.2	82.9%	2.1Meta release chart for Terminal-Bench 2.1; the chart labels each agent scaffold separately	Public reference · not scored provider-reported	Meta Superintelligence Labs permanent refresh source ↗Observed 2026-08-05 · checked 2026-08-15
Muse Spark 1.2 (xhigh) with Muse Code muse-spark-1-2-xhigh	Muse Spark 1.2	82.9%	2.1Muse Code evaluation on all 89 Terminal-Bench 2.1 tasks in isolated Daytona sandboxes	Public reference · not scored provider-reported	Muse Spark 1.2 Evaluation Methodology ↗Observed 2026-08-05 · checked 2026-08-05
GPT-5.6 Terra (reasoning configuration not stated in chart) gpt-5-6-terra-meta-muse-12-release-unspecified	GPT-5.6 Terra	81.8%	2.1Meta release chart for Terminal-Bench 2.1; the chart labels each agent scaffold separately	Relative comparison · not standard direct provider-reported	Meta Superintelligence Labs permanent refresh source ↗Observed 2026-08-05 · checked 2026-08-15
Grok 4.5 (reasoning configuration not stated in chart) grok-4-5-meta-muse-12-release-unspecified	Grok 4.5	81.6%	2.1Meta release chart for Terminal-Bench 2.1; the chart labels each agent scaffold separately	Relative comparison · not standard direct provider-reported	Meta Superintelligence Labs permanent refresh source ↗Observed 2026-08-05 · checked 2026-08-15
Gemini 3.6 Flash (reasoning configuration not stated in chart) gemini-3-6-flash-meta-muse-12-release-unspecified	Gemini 3.6 Flash	78.9%	2.1Meta release chart for Terminal-Bench 2.1; the chart labels each agent scaffold separately	Relative comparison · not standard direct provider-reported	Meta Superintelligence Labs permanent refresh source ↗Observed 2026-08-05 · checked 2026-08-15
Muse Spark 1.1 (reasoning configuration not stated in chart) muse-spark-1-1-meta-muse-12-release-unspecified	Muse Spark 1.1	76.2%	2.1Meta release chart for Terminal-Bench 2.1; the chart labels each agent scaffold separately	Relative comparison · not standard direct provider-reported	Meta Superintelligence Labs permanent refresh source ↗Observed 2026-08-05 · checked 2026-08-15
Qwen3.8 Max (xhigh, default reasoning effort) qwen-3-8-max-xhigh	Qwen3.8 Max	86.6%	2.1Claude Code on Terminal-Bench 2.1	Public reference · not scored provider-reported	Qwen3.8 Max official release ↗Observed 2026-08-03 · checked 2026-08-05
DeepSeek Harness Minimal Mode; max effort; top_p=0.95; temperature=1.0 deepseek-v4-flash-0731-max-harness	DeepSeek V4 Flash 0731	82.7%	2.1DeepSeek Harness Minimal Mode	Public reference · not scored provider-reported	DeepSeek-V4-Flash-0731 official model card and provider evaluation ↗Observed 2026-07-31 · checked 2026-08-24
Exact BenchLM registry variant GPT-5.6 Sol; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.6 Sol	91.9%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Kimi K3; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Kimi K3	88.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Claude Mythos 5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Mythos 5	88%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant GPT-5.6 Terra; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.6 Terra	87.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant GPT-5.6 Luna; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.6 Luna	84.7%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Claude Fable 5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Fable 5	84.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Grok 4.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Grok 4.5	83.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Sakana Fugu-Ultra; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Sakana Fugu-Ultra	82.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant GPT-5.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.5	82%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant SWE-1.7; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	SWE-1.7	81.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant GLM-5.2; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GLM-5.2	81%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Claude Sonnet 5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Sonnet 5	80.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Sakana Fugu; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Sakana Fugu	80.2%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Muse Spark 1.1; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Muse Spark 1.1	80%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Ornith-1.0-397B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Ornith-1.0-397B	77.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant GPT-5.3 Codex; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.3-Codex	77.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Gemini 3.5 Flash; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Gemini 3.5 Flash	76.2%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant GPT-5.4; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.4	75.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Claude Opus 4.8; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Opus 4.8	74.6%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Qwen3.7 Plus; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.7-Plus	70.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Laguna S 2.1; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Laguna S 2.1	70.2%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Qwen3.7 Max; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.7-Max	69.7%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Claude Opus 4.7 (Adaptive); bulk export does not retain a complete upstream harness configuration. claude-opus-4-7-max	Claude Opus 4.7	69.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Composer 2.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Composer 2.5	69.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant MiMo-V2.5-Pro; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MiMo-V2.5-Pro	68.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant DeepSeek V4 Pro (Max); bulk export does not retain a complete upstream harness configuration. deepseek-v4-pro-max	DeepSeek V4 Pro	67.9%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Kimi K2.6; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Kimi K2.6	66.7%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant MiniMax M3; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MiniMax M3	66%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant MiMo-V2.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MiMo-V2.5	65.8%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Claude Opus 4.6; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Opus 4.6	65.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Qwen 3.6 Max (preview); bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen 3.6 Max (preview)	65.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Ornith-1.0-35B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Ornith-1.0-35B	64.2%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Inkling; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Inkling	63.8%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant GLM-5.1; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GLM-5.1	63.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant DeepSeek V4 Pro (High); bulk export does not retain a complete upstream harness configuration. deepseek-v4-pro-high	DeepSeek V4 Pro	63.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Composer 2; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Composer 2	61.7%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Qwen3.6 Plus; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.6 Plus	61.6%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant GPT-5.4 mini; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.4 mini	60%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Step 3.7 Flash; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Step 3.7 Flash	59.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Claude Opus 4.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Opus 4.5	59.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Qwen3.6-27B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.6 27B	59.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Claude Sonnet 4.6; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Sonnet 4.6	59.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant DeepSeek V4 Pro; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	DeepSeek V4 Pro	59.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Muse Spark; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Muse Spark	59%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant MiniMax M2.7; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MiniMax M2.7	57%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant DeepSeek V4 Flash (Max); bulk export does not retain a complete upstream harness configuration. deepseek-v4-flash-max	DeepSeek V4 Flash	56.9%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant DeepSeek V4 Flash (High); bulk export does not retain a complete upstream harness configuration. deepseek-v4-flash-high	DeepSeek V4 Flash	56.6%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Nemotron 3 Ultra; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Nemotron 3 Ultra	56.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant GLM-5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GLM-5	56.2%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Hy3 Preview; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Hy3 Preview	54.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Gemini 3.5 Flash-Lite; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Gemini 3.5 Flash-Lite	54%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Qwen3.5 397B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.5 397B A17B	52.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Qwen3.6-35B-A3B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.6-35B-A3B	51.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Kimi K2.5 (Reasoning); bulk export does not retain a complete upstream harness configuration. kimi-k2-5-thinking	Kimi K2.5	50.8%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Kimi K2.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Kimi K2.5	50.8%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Claude Sonnet 4.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Sonnet 4.5	50%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Qwen3.5-122B-A10B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.5-122B-A10B	49.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant DeepSeek V4 Flash; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	DeepSeek V4 Flash	49.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Grok 4.20; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Grok 4.20	47.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant GPT-5.4 nano; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.4 nano	46.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant MAI-Thinking-1; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MAI-Thinking-1	46%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Laguna M.1; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Laguna M.1	45.8%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Ornith-1.0-9B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Ornith-1.0-9B	43.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Qwen3.5-27B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.5 27B	41.6%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant GLM-4.7; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GLM-4.7	41%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Qwen3.5-35B-A3B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.5-35B-A3B	40.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant Laguna XS.2; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Laguna XS.2	35.7%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27
Exact BenchLM registry variant GPT-5.6 Sol; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.6 Sol	91.9%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Kimi K3; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Kimi K3	88.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Claude Mythos 5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Mythos 5	88%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant GPT-5.6 Terra; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.6 Terra	87.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant GPT-5.6 Luna; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.6 Luna	84.7%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Terminus-2; as published in Gemini 3.6 Flash evaluation table. source label without registered configuration ID	GPT-5.6 Luna	84.7%	2.1Terminus 2 harness	Public reference · not scored provider-reported	Gemini 3.6 Flash product page with evaluation table ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Claude Fable 5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Fable 5	84.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Grok 4.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Grok 4.5	83.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Terminus-2; as published in Gemini 3.6 Flash evaluation table. source label without registered configuration ID	Grok 4.5	83.3%	2.1Terminus 2 harness	Public reference · not scored provider-reported	Gemini 3.6 Flash product page with evaluation table ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Sakana Fugu-Ultra; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Sakana Fugu-Ultra	82.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant GPT-5.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.5	82%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant SWE-1.7; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	SWE-1.7	81.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant GLM-5.2; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GLM-5.2	81%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Claude Sonnet 5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Sonnet 5	80.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Terminus-2; as published in Gemini 3.6 Flash evaluation table. source label without registered configuration ID	Claude Sonnet 5	80.4%	2.1Terminus 2 harness	Public reference · not scored provider-reported	Gemini 3.6 Flash product page with evaluation table ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Sakana Fugu; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Sakana Fugu	80.2%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Muse Spark 1.1; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Muse Spark 1.1	80%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
gemini-3.6-flash via Gemini API; default sampling; single-attempt / pass@1 unless noted in Google eval methodology (July 2026). Terminus-2 harness; Terminal-Bench 2.1. source label without registered configuration ID	Gemini 3.6 Flash	78%	2.1Terminus 2 harness	Score input provider-reported	Gemini 3.6 Flash product page with evaluation table ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Ornith-1.0-397B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Ornith-1.0-397B	77.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Artificial Analysis Terminal-Bench v2.1 independent evaluation (as listed on AA / BenchLM). source label without registered configuration ID	Gemini 3.6 Flash	77.5%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Gemini 3.6 Flash (Artificial Analysis) ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant GPT-5.3 Codex; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.3-Codex	77.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Gemini 3.5 Flash; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Gemini 3.5 Flash	76.2%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant GPT-5.4; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.4	75.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Claude Opus 4.8; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Opus 4.8	74.6%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Terminus-2; as published in Gemini 3.6 Flash evaluation table for Gemini 3.1 Pro. source label without registered configuration ID	Gemini 3.1 Pro Preview	73.8%	2.1Terminus 2 harness	Public reference · not scored provider-reported	Gemini 3.6 Flash product page with evaluation table ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Qwen3.7 Plus; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.7-Plus	70.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Laguna S 2.1; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Laguna S 2.1	70.2%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Qwen3.7 Max; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.7-Max	69.7%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Claude Opus 4.7 (Adaptive); bulk export does not retain a complete upstream harness configuration. claude-opus-4-7-max	Claude Opus 4.7	69.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Composer 2.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Composer 2.5	69.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant MiMo-V2.5-Pro; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MiMo-V2.5-Pro	68.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant DeepSeek V4 Pro (Max); bulk export does not retain a complete upstream harness configuration. deepseek-v4-pro-max	DeepSeek V4 Pro	67.9%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Kimi K2.6; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Kimi K2.6	66.7%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant MiniMax M3; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MiniMax M3	66%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant MiMo-V2.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MiMo-V2.5	65.8%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Claude Opus 4.6; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Opus 4.6	65.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Qwen 3.6 Max (preview); bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen 3.6 Max (preview)	65.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Ornith-1.0-35B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Ornith-1.0-35B	64.2%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Inkling; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Inkling	63.8%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant GLM-5.1; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GLM-5.1	63.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant DeepSeek V4 Pro (High); bulk export does not retain a complete upstream harness configuration. deepseek-v4-pro-high	DeepSeek V4 Pro	63.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Composer 2; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Composer 2	61.7%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Qwen3.6 Plus; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.6 Plus	61.6%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant GPT-5.4 mini; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.4 mini	60%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Step 3.7 Flash; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Step 3.7 Flash	59.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Claude Opus 4.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Opus 4.5	59.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Qwen3.6-27B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.6 27B	59.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Claude Sonnet 4.6; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Sonnet 4.6	59.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant DeepSeek V4 Pro; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	DeepSeek V4 Pro	59.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Muse Spark; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Muse Spark	59%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant MiniMax M2.7; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MiniMax M2.7	57%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant DeepSeek V4 Flash (Max); bulk export does not retain a complete upstream harness configuration. deepseek-v4-flash-max	DeepSeek V4 Flash	56.9%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant DeepSeek V4 Flash (High); bulk export does not retain a complete upstream harness configuration. deepseek-v4-flash-high	DeepSeek V4 Flash	56.6%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Nemotron 3 Ultra; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Nemotron 3 Ultra	56.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant GLM-5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GLM-5	56.2%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Hy3 Preview; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Hy3 Preview	54.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Gemini 3.5 Flash-Lite; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Gemini 3.5 Flash-Lite	54%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
gemini-3.5-flash-lite via Gemini API; configuration as stated in the July 2026 Google launch materials. Terminal-Bench 2.1 as published in Google launch blog. source label without registered configuration ID	Gemini 3.5 Flash-Lite	54%	2.1Provider evaluation harness	Public reference · not scored provider-reported	Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Qwen3.5 397B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.5 397B A17B	52.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Qwen3.6-35B-A3B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.6-35B-A3B	51.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Kimi K2.5 (Reasoning); bulk export does not retain a complete upstream harness configuration. kimi-k2-5-thinking	Kimi K2.5	50.8%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Kimi K2.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Kimi K2.5	50.8%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Claude Sonnet 4.5; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Claude Sonnet 4.5	50%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Qwen3.5-122B-A10B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.5-122B-A10B	49.4%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant DeepSeek V4 Flash; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	DeepSeek V4 Flash	49.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Grok 4.20; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Grok 4.20	47.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant GPT-5.4 nano; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GPT-5.4 nano	46.3%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant MAI-Thinking-1; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MAI-Thinking-1	46%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Laguna M.1; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Laguna M.1	45.8%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Ornith-1.0-9B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Ornith-1.0-9B	43.1%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Qwen3.5-27B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.5 27B	41.6%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant GLM-4.7; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	GLM-4.7	41%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Qwen3.5-35B-A3B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.5-35B-A3B	40.5%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
Exact BenchLM registry variant Laguna XS.2; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Laguna XS.2	35.7%	2026Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21
As published in Gemini 3.5 Flash-Lite launch materials (prior Flash-Lite comparison). source label without registered configuration ID	Gemini 3.1 Flash-Lite	31%	2.1Provider evaluation harness	Public reference · not scored provider-reported	Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber ↗Observed 2026-07-21 · checked 2026-07-21
Exact model variant as listed on the BenchLM public terminal-bench leaderboard page (verified 2026-07-21). source label without registered configuration ID	Claude Opus 4.7	69.4%	2.0/2.1BenchLM aggregated public evaluation	Score input source-checked	Terminal-Bench 2.0 Leaderboard & Scores — July 2026 ↗Observed 2026-07-20 · checked 2026-07-21
Exact model variant as listed on the BenchLM public terminal-bench leaderboard page (verified 2026-07-21). source label without registered configuration ID	Qwen3.6 Max	65.4%	2.0/2.1BenchLM aggregated public evaluation	Score input source-checked	Terminal-Bench 2.0 Leaderboard & Scores — July 2026 ↗Observed 2026-07-20 · checked 2026-07-21
Exact model variant as listed on the BenchLM public terminal-bench leaderboard page (verified 2026-07-21). source label without registered configuration ID	Hy3	54.4%	2.0/2.1BenchLM aggregated public evaluation	Score input source-checked	Terminal-Bench 2.0 Leaderboard & Scores — July 2026 ↗Observed 2026-07-20 · checked 2026-07-21
KimiCode harness; reasoning max; Terminal-Bench 2.1 as published in the Kimi K3 launch blog. kimi-k3-max	Kimi K3	88.3%	2.1KimiCode	Public reference · not scored provider-reported	Kimi K3: Open Frontier Intelligence ↗Observed 2026-07-16 · checked 2026-07-21
Muse Spark 1.1 (xhigh); Artificial Analysis independent evaluation configuration as published. muse-spark-1-1-xhigh	Muse Spark 1.1	77.9%	2.1Artificial Analysis independent evaluation	Public reference · not scored source-checked	Muse Spark 1.1 (xhigh) analysis ↗Observed 2026-07-16 · checked 2026-07-16
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	GPT-5.6 Sol	91.9%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	GPT-5.6 Sol	88.0%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for gpt-5-6-sol ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	GPT-5.6 Terra	88.0%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for gpt-5-6-terra ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	Claude Mythos 5	88%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	GPT-5.6 Terra	87.4%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	GPT-5.6 Luna	84.7%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	Claude Fable 5	84.6%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for claude-fable-5 ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	Claude Opus 4.8	84.6%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for claude-opus-4-8 ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	Claude Fable 5	84.3%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	GPT-5.5	84.3%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for gpt-5-5 ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	Grok 4.5	83.3%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	GPT-5.5	82%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	Grok 4.5	81.7%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for grok-4-5 ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	GLM-5.2	81%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	GPT-5.6 Luna	80.9%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for gpt-5-6-luna ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	Claude Sonnet 5	80.5%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for claude-sonnet-5 ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	Claude Sonnet 5	80.4%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	Muse Spark 1.1	80%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	GPT-5.4	78.3%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for gpt-5-4 ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	Gemini 3.5 Flash	76.2%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	Claude Opus 4.8	74.6%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	Qwen3.7-Max	74.5%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for qwen3-7-max ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	Gemini 3.1 Pro Preview	73.8%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for gemini-3-1-pro-preview ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	Qwen3.7-Plus	70.3%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	Qwen3.7-Max	69.7%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	MiMo-V2.5-Pro	68.4%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Kimi K2.7 Code source label without registered configuration ID	Kimi K2.7 Code	67.4%	2.1Artificial Analysis independent evaluation	Public reference · not scored source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	MiniMax M3	66%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Kimi K2.6 source label without registered configuration ID	Kimi K2.6	65.9%	2.1Artificial Analysis independent evaluation	Public reference · not scored source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	MiniMax M3	65.2%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for minimax-m3 ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	DeepSeek V4 Pro	64.0%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for deepseek-v4-pro ↗Observed 2026-07-15 · checked 2026-07-15
MiMo-V2.5 source label without registered configuration ID	MiMo-V2.5	63.7%	2.1Artificial Analysis independent evaluation	Public reference · not scored source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	DeepSeek V4 Flash	61.8%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for deepseek-v4-flash ↗Observed 2026-07-15 · checked 2026-07-15
Qwen3.6 27B (Reasoning) source label without registered configuration ID	Qwen3.6 27B	60.7%	2.1Artificial Analysis independent evaluation	Public reference · not scored source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
GPT-5.4 mini (xhigh) gpt-5-4-mini-xhigh	GPT-5.4 mini	59.2%	2.1Artificial Analysis independent evaluation	Public reference · not scored source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	DeepSeek V4 Pro	59.1%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Claude 4.5 Sonnet (Reasoning) source label without registered configuration ID	Claude Sonnet 4.5	55.8%	2.1Artificial Analysis independent evaluation	Score input source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Nemotron 3 Ultra 550B A55B (Reasoning) source label without registered configuration ID	Nemotron 3 Ultra	53.9%	2.1Artificial Analysis independent evaluation	Score input source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Qwen3.5 397B A17B (Reasoning) source label without registered configuration ID	Qwen3.5 397B A17B	51.3%	2.1Artificial Analysis independent evaluation	Score input source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	Mistral Medium 3.5	50.6%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for mistral-medium-3-5 ↗Observed 2026-07-15 · checked 2026-07-15
GLM-4.6 (Reasoning) source label without registered configuration ID	GLM-4.6	49.4%	2.1Artificial Analysis independent evaluation	Score input source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Exact model variant as listed on the BenchLM public benchmark leaderboard page. source label without registered configuration ID	DeepSeek V4 Flash	49.1%	2.1BenchLM aggregated public evaluation	Source carrier · no additional scoring weight source-checked	BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15
Qwen3.5 122B A10B (Reasoning) source label without registered configuration ID	Qwen3.5 122B	47.6%	2.1Artificial Analysis independent evaluation	Score input source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
DeepSeek V3.2 (Reasoning) source label without registered configuration ID	DeepSeek V3.2	46.8%	2.1Artificial Analysis independent evaluation	Score input source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
GLM-4.7 (Reasoning) source label without registered configuration ID	GLM-4.7	45.3%	2.1Artificial Analysis independent evaluation	Public reference · not scored source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
DeepSeek V3.1 Terminus (Reasoning) source label without registered configuration ID	DeepSeek V3.1 Terminus	44.9%	2.1Artificial Analysis independent evaluation	Score input source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	Grok 4.3	39.7%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for grok-4-3 ↗Observed 2026-07-15 · checked 2026-07-15
Gemma 4 26B A4B (Reasoning) source label without registered configuration ID	Gemma 4 26B	39.0%	2.1Artificial Analysis independent evaluation	Score input source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
NVIDIA Nemotron 3 Super 120B A12B (Reasoning) source label without registered configuration ID	Nemotron 3 Super	38.6%	2.1Artificial Analysis independent evaluation	Score input source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Claude 4 Sonnet (Reasoning) source label without registered configuration ID	Claude Sonnet 4	36.3%	2.1Artificial Analysis independent evaluation	Score input source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
GPT-5 (high) gpt-5-high	GPT-5	35.2%	2.1Artificial Analysis independent evaluation	Public reference · not scored source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Nova 2.0 Pro Preview (medium) source label without registered configuration ID	Nova 2 Pro	29.6%	2.1Artificial Analysis independent evaluation	Score input source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Gemini 2.5 Pro source label without registered configuration ID	Gemini 2.5 Pro	28.5%	2.1Artificial Analysis independent evaluation	Public reference · not scored source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Command A+ source label without registered configuration ID	Command A+	22.8%	2.1Artificial Analysis independent evaluation	Public reference · not scored source-checked	Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	Mistral Small 4	21.0%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for mistral-small-4 ↗Observed 2026-07-15 · checked 2026-07-15
Artificial Analysis public model evaluation page; exact provider variant as listed on the page. source label without registered configuration ID	Mistral Large 3	12.0%	2.1Artificial Analysis evaluation harness	Public reference · not scored source-checked	Artificial Analysis evaluations for mistral-large-3 ↗Observed 2026-07-15 · checked 2026-07-15
Codex; reasoning max; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	GPT-5.6 Terra	78.4%	2.1Codex	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-11 · checked 2026-08-16
Codex; reasoning max; Terminal-Bench 2.1 verified submission. gpt-5-6-terra-max	GPT-5.6 Terra	78.4%	2.1Codex	Public reference · not scored official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-11 · checked 2026-07-15
Codex; reasoning max; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	GPT-5.6 Luna	75.7%	2.1Codex	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-11 · checked 2026-08-16
Codex; reasoning max; Terminal-Bench 2.1 verified submission. gpt-5-6-luna-max	GPT-5.6 Luna	75.7%	2.1Codex	Public reference · not scored official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-11 · checked 2026-07-15
Muse Spark 1.1; provider evaluation configuration as published in the Meta Muse Spark 1.1 evaluation report (Terminal-Bench 2.0). muse-spark-1-1-xhigh	Muse Spark 1.1	80%	2.1Provider evaluation harness	Public reference · not scored provider-reported	Meta AI Muse Spark 1.1 evaluation report ↗Observed 2026-07-09 · checked 2026-07-16
Cursor CLI; reasoning high; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	Grok 4.5	79.3%	2.1Cursor CLI	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-08-16
Claude Code; reasoning high; Terminal-Bench 2.1 verified submission. claude-opus-4-8-high	Claude Opus 4.8	78.9%	2.1Claude Code	Public reference · not scored official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-07-15
Claude Code; reasoning high; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	Claude Opus 4.8	78.9%	2.1Claude Code	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-08-16
mini-SWE-agent; reasoning xhigh; Terminal-Bench 2.1 verified submission. muse-spark-1-1-xhigh	Muse Spark 1.1	76.2%	2.1mini-SWE-agent	Public reference · not scored official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-07-16
mini-SWE-agent; reasoning xhigh; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	Muse Spark 1.1	76.2%	2.1mini-SWE-agent	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-08-16
Claude Code; reasoning high; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	Claude Sonnet 5	74.6%	2.1Claude Code	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-08-16
Claude Code; reasoning high; Terminal-Bench 2.1 verified submission. claude-sonnet-5-high	Claude Sonnet 5	74.6%	2.1Claude Code	Public reference · not scored official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-07-15
LongCat-2.0 (in-house harness) longcat-2-0-default	LongCat-2.0	70.8%	2.1LongCat unified in-house harness	Score input provider-reported	LongCat-2.0 model card ↗Observed 2026-06-30 · checked 2026-08-15
Claude Code; reasoning xhigh; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	Claude Fable 5	83.8%	2.1Claude Code	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-06-07 · checked 2026-08-16
Claude Code; reasoning xhigh; Terminal-Bench 2.1 verified submission. claude-fable-5-xhigh	Claude Fable 5	83.8%	2.1Claude Code	Public reference · not scored official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-06-07 · checked 2026-07-15
Terminus 2; reasoning high; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	Claude Fable 5	80.5%	2.1Terminus 2	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-06-05 · checked 2026-08-16
Terminus 2; reasoning high; Terminal-Bench 2.1 verified submission. claude-fable-5-high	Claude Fable 5	80.4%	2.1Terminus 2	Public reference · not scored official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-06-05 · checked 2026-07-15
Terminus-2 agent system; provider-published Terminal-Bench 2.1 evaluation. gpt-5-5-high	GPT-5.5	78.2%	2.1Terminus 2	Public reference · not scored provider-reported	Gemini 3.5 Flash model card ↗Observed 2026-05-19 · checked 2026-07-15
Terminus-2 agent system; provider-published Terminal-Bench 2.1 evaluation. gemini-3-5-flash-high	Gemini 3.5 Flash	76.2%	2.1Terminus 2	Public reference · not scored provider-reported	Gemini 3.5 Flash model card ↗Observed 2026-05-19 · checked 2026-07-15
Terminus-2 agent system; provider-published Terminal-Bench 2.1 evaluation. source label without registered configuration ID	Gemini 3.1 Pro Preview	70.3%	2.1Terminus 2	Source carrier · no additional scoring weight provider-reported	Gemini 3.5 Flash model card ↗Observed 2026-05-19 · checked 2026-07-15
Codex; reasoning xhigh; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	GPT-5.5	83.2%	2.1Codex	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-08-16
Codex; reasoning xhigh; Terminal-Bench 2.1 verified submission. gpt-5-5-xhigh	GPT-5.5	83.1%	2.1Codex	Public reference · not scored official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-07-15
Terminus 2; reasoning xhigh; Terminal-Bench 2.1 verified submission. gpt-5-5-xhigh	GPT-5.5	78%	2.1Terminus 2	Public reference · not scored official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-07-15
Terminus 2; reasoning xhigh; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	GPT-5.5	78.0%	2.1Terminus 2	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-08-16
Opus 4.7; reasoning=max; agent=Claude Code claude-opus-4-7-max	Claude Opus 4.7	68.9%	2.1Claude Code	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-08-16
Opus 4.7; reasoning=max; agent=Terminus 2 claude-opus-4-7-max	Claude Opus 4.7	66.1%	2.1Terminus 2	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-08-16
Gemini CLI; reasoning high; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	Gemini 3.1 Pro Preview	65.8%	2.1Gemini CLI	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-07-15
Terminus 2; reasoning high; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	Gemini 3.1 Pro Preview	65.6%	2.1Terminus 2	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-07-15
Claude Code; reasoning max; Terminal-Bench 2.1 verified submission. source label without registered configuration ID	GLM-5.1	58.7%	2.1Claude Code	Score input official-leaderboard	Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-07-16

Exact published model / configuration

Canonical product

Score

Version / system

Evidence

Source / observed

Qwen3.8-Flash-Next (Artificial Analysis independent run)

qwen-3-8-flash-next-aa-unspecified

Qwen3.8-Flash-Next

86.1%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Public reference · not scored

independently-verified

Qwen3.8-Flash-Next individual evaluations ↗Observed 2026-08-29 · checked 2026-08-29

Claude Opus 4.7 (Adaptive Reasoning, Max Effort) (Artificial Analysis independent run)

claude-opus-4-7-max

83.1%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Claude Opus 4.7 (Adaptive Reasoning, Max Effort) individual evaluations ↗Observed 2026-08-29 · checked 2026-08-29

GLM-5.2 (max) (Artificial Analysis independent run)

glm-5-2-max

77.9%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

GLM-5.2 (max) individual evaluations ↗Observed 2026-08-29 · checked 2026-08-29

Qwen3.5 397B A17B (Reasoning) (Artificial Analysis completed independent run)

qwen3-5-397b-thinking

51.3%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Qwen3.5 397B A17B (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

Mistral Medium 3.5 (Artificial Analysis completed independent run)

mistral-medium-3-5-128b-aa-reasoning-default

Mistral Medium 3.5 128B

50.6%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Mistral Medium 3.5 current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

LongCat 2.0 (Artificial Analysis independent run)

longcat-2-0-default

LongCat-2.0

50.2%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

LongCat 2.0 individual evaluations ↗Observed 2026-08-29 · checked 2026-08-29

Qwen3.5 122B A10B (Reasoning) (Artificial Analysis completed independent run)

qwen3-5-122b-a10b-aa-reasoning-default

47.6%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Qwen3.5 122B A10B (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

Kimi K2.5 (Reasoning) (Artificial Analysis completed independent run)

kimi-k2-5-thinking

45.7%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Kimi K2.5 (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

Qwen3.6 35B A3B (Reasoning) (Artificial Analysis completed independent run)

qwen3-6-35b-a3b-aa-reasoning-default

44.9%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Qwen3.6 35B A3B (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

Gemma 4 26B A4B (Reasoning) (Artificial Analysis completed independent run)

gemma-4-26b-a4b-aa-reasoning-default

Gemma 4 26B A4B

39.0%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Gemma 4 26B A4B (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

Mistral Small 4 (Reasoning) (Artificial Analysis completed independent run)

mistral-small-4-reasoning

Mistral Small 4

21.0%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Mistral Small 4 (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

Trinity Large Thinking (Artificial Analysis completed independent run)

trinity-large-thinking-aa-reasoning-default

Trinity-Large-Thinking

20.6%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Trinity Large Thinking current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

GPT-4.1 mini (Artificial Analysis completed independent run)

gpt-4-1-mini-epoch-gpt-4-1-mini-2025-04-14

GPT-4.1 mini

10.1%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

GPT-4.1 mini current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

Llama 4 Maverick (Artificial Analysis completed independent run)

llama-4-maverick-aa-non-reasoning-default

Llama 4 Maverick

7.9%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Llama 4 Maverick current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

NVIDIA Nemotron 3 Nano 30B A3B (Reasoning) (Artificial Analysis completed independent run)

nemotron-3-nano-30b-aa-reasoning-default

Nemotron 3 Nano 30B

6.7%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

NVIDIA Nemotron 3 Nano 30B A3B (Reasoning) current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

Gemma 3 27B Instruct (Artificial Analysis completed independent run)

gemma-3-27b-aa-non-reasoning-default

Gemma 3 27B

4.5%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Gemma 3 27B Instruct current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

GPT-4.1 nano (Artificial Analysis completed independent run)

gpt-4-1-nano-epoch-gpt-4-1-nano-2025-04-14

GPT-4.1 nano

3.7%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

GPT-4.1 nano current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

Llama 4 Scout (Artificial Analysis completed independent run)

llama-4-scout-aa-non-reasoning-default

Llama 4 Scout

3.7%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Llama 4 Scout current model and individual-evaluation page ↗Observed 2026-08-29 · checked 2026-08-29

Claude Opus 5 (max; Artificial Analysis independent run)

claude-opus-5-max

Claude Opus 5

89.1%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Claude Opus 5 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

Grok 4.6 (high; Artificial Analysis independent run)

grok-4-6-high

Grok 4.6

88.4%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Grok 4.6 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

GPT-5.6 Sol (max; Artificial Analysis independent run)

gpt-5-6-sol-max

88.0%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

GPT-5.6 Sol individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

GPT-5.6 Terra (max; Artificial Analysis independent run)

gpt-5-6-terra-max

88.0%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

GPT-5.6 Terra individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

Kimi K3 (max; Artificial Analysis independent run)

kimi-k3-max

85.0%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Kimi K3 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

Claude Fable 5 (max; Artificial Analysis independent run)

claude-fable-5-max

84.6%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Claude Fable 5 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

Claude Opus 4.8 (max; Artificial Analysis independent run)

claude-opus-4-8-max

84.6%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Claude Opus 4.8 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

GLM-5.3-Flash (max; Artificial Analysis independent run)

glm-5-3-flash-max

GLM-5.3-Flash

84.3%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

GLM-5.3-Flash individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

GPT-5.5 (xhigh; Artificial Analysis independent run)

gpt-5-5-xhigh

84.3%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

GPT-5.5 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

GLM-5.3 (max; Artificial Analysis independent run)

glm-5-3-max

GLM-5.3

83.9%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

GLM-5.3 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

Grok 4.5 (high; Artificial Analysis independent run)

grok-4-5-aa-2-high

81.6%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Grok 4.5 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

GPT-5.6 Luna (max; Artificial Analysis independent run)

gpt-5-6-luna-max

80.9%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

GPT-5.6 Luna individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

Claude Sonnet 5 (max; Artificial Analysis independent run)

claude-sonnet-5-max

80.5%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Claude Sonnet 5 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

Muse Spark 1.2 (xhigh; Artificial Analysis independent run)

muse-spark-1-2-xhigh

80.1%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Muse Spark 1.2 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

DeepSeek V4 Pro 0813 (max; Artificial Analysis independent run)

deepseek-v4-pro-0813-max

78.7%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Public reference · not scored

independently-verified

DeepSeek V4 Pro 0813 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

Gemini 3.7 Flash (medium; Artificial Analysis independent run)

gemini-3-7-flash-medium

Gemini 3.7 Flash

78.3%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Gemini 3.7 Flash individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

GPT-5.4 (xhigh; Artificial Analysis independent run)

gpt-5-4-xhigh

78.3%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

GPT-5.4 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

Muse Spark 1.1 (xhigh; Artificial Analysis independent run)

muse-spark-1-1-xhigh

77.9%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Muse Spark 1.1 individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

Gemini 3.6 Flash (high; Artificial Analysis independent run)

gemini-3-6-flash-high

DeepSeek V4 Flash Vision Exp

77.5%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Score input

independently-verified

Gemini 3.6 Flash individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

DeepSeek V4 Flash Vision Exp (max; Artificial Analysis independent run)

deepseek-v4-flash-vision-exp-max-harness

74.2%

2.1Artificial Analysis Terminal-Bench v2.1 in e2b

Public reference · not scored

independently-verified

DeepSeek V4 Flash Vision Exp individual evaluations ↗Observed 2026-08-27 · checked 2026-08-27

DeepSeek Harness Minimal Mode; max effort; top_p=0.95; temperature=1.0

deepseek-v4-flash-vision-exp-max-harness

DeepSeek V4 Flash Vision Exp

83.9%

2.1DeepSeek Harness Minimal Mode

Public reference · not scored

provider-reported

DeepSeek-V4-Flash-Vision-Exp release and provider evaluation ↗Observed 2026-08-21 · checked 2026-08-24

Claude Opus 4.6 Max as published by Qwen

claude-opus-4-6-max

78.2%

2.1Qwen3.8-27B official text table

Relative comparison · not standard direct

provider-reported

Qwen3.8-27B official model card ↗Observed 2026-08-15 · checked 2026-08-15

Qwen3.8-27B (xhigh)

qwen-3-8-27b-xhigh

Qwen3.8-27B

73%

2.1Qwen3.8-27B official text table

Score input

provider-reported

Qwen3.8-27B official model card ↗Observed 2026-08-15 · checked 2026-08-15

Qwen3.7-Plus as published by Qwen

qwen-3-7-plus-unspecified

64%

2.1Qwen3.8-27B official text table

Relative comparison · not standard direct

provider-reported

Qwen3.8-27B official model card ↗Observed 2026-08-15 · checked 2026-08-15

Qwen3.6-27B as published by Qwen

qwen3-6-27b-default

63.4%

2.1Qwen3.8-27B official text table

Relative comparison · not standard direct

provider-reported

Qwen3.8-27B official model card ↗Observed 2026-08-15 · checked 2026-08-15

GPT-5.6 Sol as published by Z.AI

gpt-5-6-sol-max

88.8%

2.1Z.AI GLM-5.3 official comparison table

Relative comparison · not standard direct

provider-reported

GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15

Kimi K3 as published by Z.AI

kimi-k3-max

88.3%

2.1Z.AI GLM-5.3 official comparison table

Relative comparison · not standard direct

provider-reported

GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15

GLM-5.3 (max)

glm-5-3-max

GLM-5.3

88.2%

2.1Z.AI GLM-5.3 official comparison table

Score input

provider-reported

GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15

Claude Fable 5 w/ fallback as published by Z.AI

claude-fable-5-max

88%

2.1Z.AI GLM-5.3 official comparison table

Relative comparison · not standard direct

provider-reported

GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15

DeepSeek V4 Pro 0813 as published by Z.AI

deepseek-v4-pro-0813-max

87.9%

2.1Z.AI GLM-5.3 official comparison table

Relative comparison · not standard direct

provider-reported

GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15

Qwen3.8 Max as published by Z.AI

qwen-3-8-max-xhigh

Qwen3.8 Max

86.6%

2.1Z.AI GLM-5.3 official comparison table

Relative comparison · not standard direct

provider-reported

GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15

Claude Opus 4.8 as published by Z.AI

claude-opus-4-8-max

85%

2.1Z.AI GLM-5.3 official comparison table

Relative comparison · not standard direct

provider-reported

GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15

GLM-5.2 as published by Z.AI

glm-5-2-max

81%

2.1Z.AI GLM-5.3 official comparison table

Relative comparison · not standard direct

provider-reported

GLM-5.3: Frontier Coding with Emergent Cyber Capabilities ↗Observed 2026-08-14 · checked 2026-08-15

Kimi K3

kimi-k3-deepseek-0813-release-unspecified

88.3%

2.1DeepSeek V4 Pro 0813 official GA comparison table

Public reference · not scored

provider-reported

DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

Fable 5 (with fallback)

claude-fable-5-deepseek-0813-release-unspecified

88%

2.1DeepSeek V4 Pro 0813 official GA comparison table

Public reference · not scored

provider-reported

DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

DeepSeek V4 Pro 0813

deepseek-v4-pro-0813-deepseek-0813-release-unspecified

87.9%

2.1DeepSeek V4 Pro 0813 official GA comparison table

Public reference · not scored

provider-reported

DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

GPT-5.6 Terra

gpt-5-6-terra-max

87.4%

2.1Terminal-Bench 2.1 default Terminus 2 agent harness

Public reference · not scored

provider-reported

Google DeepMind permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

Gemini 3.7 Flash

gemini-3-7-flash-medium

Gemini 3.7 Flash

85.8%

2.1Terminal-Bench 2.1 default Terminus 2 agent harness

Score input

provider-reported

Google DeepMind permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

Claude Opus 4.8

claude-opus-4-8-deepseek-0813-release-unspecified

85%

2.1DeepSeek V4 Pro 0813 official GA comparison table

Public reference · not scored

provider-reported

DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

Muse Spark 1.2

muse-spark-1-2-xhigh

82.9%

2.1Terminal-Bench 2.1 default Terminus 2 agent harness

Public reference · not scored

provider-reported

Google DeepMind permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

DeepSeek V4 Flash 0731

deepseek-v4-flash-0731-deepseek-0813-release-unspecified

DeepSeek V4 Flash 0731

82.7%

2.1DeepSeek V4 Pro 0813 official GA comparison table

Source carrier · no additional scoring weight

provider-reported

DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

GLM-5.2

glm-5-2-deepseek-0813-release-unspecified

81%

2.1DeepSeek V4 Pro 0813 official GA comparison table

Public reference · not scored

provider-reported

DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

Claude Sonnet 5

claude-sonnet-5-max

80.4%

2.1Terminal-Bench 2.1 default Terminus 2 agent harness

Public reference · not scored

provider-reported

Google DeepMind permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

Gemini 3.6 Flash

gemini-3-6-flash-high

78%

2.1Terminal-Bench 2.1 default Terminus 2 agent harness

Public reference · not scored

provider-reported

Google DeepMind permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

DeepSeek V4 Pro Preview

deepseek-v4-pro-deepseek-0813-release-unspecified

72.1%

2.1DeepSeek V4 Pro 0813 official GA comparison table

Public reference · not scored

provider-reported

DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

DeepSeek V4 Flash Preview

deepseek-v4-flash-deepseek-0813-release-unspecified

NVIDIA Nemotron 3.5 Lightning 30B-A3B

61.8%

2.1DeepSeek V4 Pro 0813 official GA comparison table

Public reference · not scored

provider-reported

DeepSeek permanent refresh source ↗Observed 2026-08-13 · checked 2026-08-13

NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16; NVIDIA release evaluation; temperature=1.0; top_p=0.95

nemotron-3-5-lightning-30b-a3b-default

24.6%

2.1NVIDIA NeMo Gym / NeMo Evaluator SDK consistent release harness

Score input

provider-reported

NVIDIA Nemotron 3.5 Lightning 30B-A3B BF16 model card ↗Observed 2026-08-11 · checked 2026-08-12

Muse Glimmer-30B; high reasoning; temperature=1.0; top_p=0.95; top_k=64

muse-glimmer-30b-high

Muse Glimmer 30B

51.7%

2.1Artificial Analysis Terminal-Bench 2.1 evaluation in an E2B sandbox

Public reference · not scored

provider-reported

Muse Glimmer Evaluation Methodology ↗Observed 2026-08-10 · checked 2026-08-10

Claude Opus 5 (reasoning configuration not stated in chart)

claude-opus-5-meta-muse-12-release-unspecified

Claude Opus 5

86.7%

2.1Meta release chart for Terminal-Bench 2.1; the chart labels each agent scaffold separately

Relative comparison · not standard direct

provider-reported

Meta Superintelligence Labs permanent refresh source ↗Observed 2026-08-05 · checked 2026-08-15

Muse Spark 1.2 (xhigh)

muse-spark-1-2-xhigh

82.9%

2.1Meta release chart for Terminal-Bench 2.1; the chart labels each agent scaffold separately

Public reference · not scored

provider-reported

Meta Superintelligence Labs permanent refresh source ↗Observed 2026-08-05 · checked 2026-08-15

Muse Spark 1.2 (xhigh) with Muse Code

muse-spark-1-2-xhigh

82.9%

2.1Muse Code evaluation on all 89 Terminal-Bench 2.1 tasks in isolated Daytona sandboxes

Public reference · not scored

provider-reported

Muse Spark 1.2 Evaluation Methodology ↗Observed 2026-08-05 · checked 2026-08-05

GPT-5.6 Terra (reasoning configuration not stated in chart)

gpt-5-6-terra-meta-muse-12-release-unspecified

81.8%

2.1Meta release chart for Terminal-Bench 2.1; the chart labels each agent scaffold separately

Relative comparison · not standard direct

provider-reported

Meta Superintelligence Labs permanent refresh source ↗Observed 2026-08-05 · checked 2026-08-15

Grok 4.5 (reasoning configuration not stated in chart)

grok-4-5-meta-muse-12-release-unspecified

81.6%

2.1Meta release chart for Terminal-Bench 2.1; the chart labels each agent scaffold separately

Relative comparison · not standard direct

provider-reported

Meta Superintelligence Labs permanent refresh source ↗Observed 2026-08-05 · checked 2026-08-15

Gemini 3.6 Flash (reasoning configuration not stated in chart)

gemini-3-6-flash-meta-muse-12-release-unspecified

78.9%

2.1Meta release chart for Terminal-Bench 2.1; the chart labels each agent scaffold separately

Relative comparison · not standard direct

provider-reported

Meta Superintelligence Labs permanent refresh source ↗Observed 2026-08-05 · checked 2026-08-15

Muse Spark 1.1 (reasoning configuration not stated in chart)

muse-spark-1-1-meta-muse-12-release-unspecified

76.2%

2.1Meta release chart for Terminal-Bench 2.1; the chart labels each agent scaffold separately

Relative comparison · not standard direct

provider-reported

Meta Superintelligence Labs permanent refresh source ↗Observed 2026-08-05 · checked 2026-08-15

Qwen3.8 Max (xhigh, default reasoning effort)

qwen-3-8-max-xhigh

Qwen3.8 Max

86.6%

2.1Claude Code on Terminal-Bench 2.1

Public reference · not scored

provider-reported

Qwen3.8 Max official release ↗Observed 2026-08-03 · checked 2026-08-05

DeepSeek Harness Minimal Mode; max effort; top_p=0.95; temperature=1.0

deepseek-v4-flash-0731-max-harness

DeepSeek V4 Flash 0731

82.7%

2.1DeepSeek Harness Minimal Mode

Public reference · not scored

provider-reported

DeepSeek-V4-Flash-0731 official model card and provider evaluation ↗Observed 2026-07-31 · checked 2026-08-24

Exact BenchLM registry variant GPT-5.6 Sol; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

91.9%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Kimi K3; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

88.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Claude Mythos 5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

88%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant GPT-5.6 Terra; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

87.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant GPT-5.6 Luna; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

84.7%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Claude Fable 5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

84.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Grok 4.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

83.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Sakana Fugu-Ultra; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Sakana Fugu-Ultra

82.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant GPT-5.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

82%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant SWE-1.7; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

SWE-1.7

81.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant GLM-5.2; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

81%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Claude Sonnet 5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

80.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Sakana Fugu; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Sakana Fugu

80.2%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Muse Spark 1.1; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

80%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Ornith-1.0-397B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Ornith-1.0-397B

77.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant GPT-5.3 Codex; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

GPT-5.3-Codex

77.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Gemini 3.5 Flash; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

76.2%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant GPT-5.4; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

75.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Claude Opus 4.8; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

74.6%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Qwen3.7 Plus; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

70.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Laguna S 2.1; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Laguna S 2.1

70.2%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Qwen3.7 Max; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

69.7%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Claude Opus 4.7 (Adaptive); bulk export does not retain a complete upstream harness configuration.

claude-opus-4-7-max

69.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Composer 2.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Composer 2.5

69.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant MiMo-V2.5-Pro; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

68.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant DeepSeek V4 Pro (Max); bulk export does not retain a complete upstream harness configuration.

deepseek-v4-pro-max

67.9%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Kimi K2.6; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

66.7%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant MiniMax M3; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

66%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant MiMo-V2.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

65.8%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Claude Opus 4.6; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

65.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Qwen 3.6 Max (preview); bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Qwen 3.6 Max (preview)

65.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Ornith-1.0-35B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Ornith-1.0-35B

64.2%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Inkling; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Inkling

63.8%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant GLM-5.1; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

63.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant DeepSeek V4 Pro (High); bulk export does not retain a complete upstream harness configuration.

deepseek-v4-pro-high

63.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Composer 2; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Composer 2

61.7%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Qwen3.6 Plus; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Qwen3.6 Plus

61.6%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant GPT-5.4 mini; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

60%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Step 3.7 Flash; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Step 3.7 Flash

59.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Claude Opus 4.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Claude Opus 4.5

59.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Qwen3.6-27B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

59.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Claude Sonnet 4.6; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Claude Sonnet 4.6

59.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant DeepSeek V4 Pro; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

59.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Muse Spark; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Muse Spark

59%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant MiniMax M2.7; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

MiniMax M2.7

57%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant DeepSeek V4 Flash (Max); bulk export does not retain a complete upstream harness configuration.

deepseek-v4-flash-max

56.9%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant DeepSeek V4 Flash (High); bulk export does not retain a complete upstream harness configuration.

deepseek-v4-flash-high

56.6%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Nemotron 3 Ultra; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

56.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant GLM-5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

GLM-5

56.2%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Hy3 Preview; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Hy3 Preview

54.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Gemini 3.5 Flash-Lite; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

54%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Qwen3.5 397B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

52.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Qwen3.6-35B-A3B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

51.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Kimi K2.5 (Reasoning); bulk export does not retain a complete upstream harness configuration.

kimi-k2-5-thinking

50.8%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Kimi K2.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

50.8%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Claude Sonnet 4.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

50%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Qwen3.5-122B-A10B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

49.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant DeepSeek V4 Flash; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

49.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Grok 4.20; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Grok 4.20

47.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant GPT-5.4 nano; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

GPT-5.4 nano

46.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant MAI-Thinking-1; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

MAI-Thinking-1

46%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Laguna M.1; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Laguna M.1

45.8%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Ornith-1.0-9B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Ornith-1.0-9B

43.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Qwen3.5-27B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Qwen3.5 27B

41.6%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant GLM-4.7; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

41%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Qwen3.5-35B-A3B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Qwen3.5-35B-A3B

40.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant Laguna XS.2; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Laguna XS.2

35.7%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 2026-07-27 ↗Observed 2026-07-27 · checked 2026-07-27

Exact BenchLM registry variant GPT-5.6 Sol; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

91.9%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Kimi K3; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

88.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Claude Mythos 5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

88%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant GPT-5.6 Terra; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

87.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant GPT-5.6 Luna; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

84.7%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Terminus-2; as published in Gemini 3.6 Flash evaluation table.

source label without registered configuration ID

84.7%

2.1Terminus 2 harness

Public reference · not scored

provider-reported

Gemini 3.6 Flash product page with evaluation table ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Claude Fable 5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

84.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Grok 4.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

83.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Terminus-2; as published in Gemini 3.6 Flash evaluation table.

source label without registered configuration ID

83.3%

2.1Terminus 2 harness

Public reference · not scored

provider-reported

Gemini 3.6 Flash product page with evaluation table ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Sakana Fugu-Ultra; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Sakana Fugu-Ultra

82.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant GPT-5.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

82%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant SWE-1.7; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

SWE-1.7

81.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant GLM-5.2; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

81%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Claude Sonnet 5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

80.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Terminus-2; as published in Gemini 3.6 Flash evaluation table.

source label without registered configuration ID

80.4%

2.1Terminus 2 harness

Public reference · not scored

provider-reported

Gemini 3.6 Flash product page with evaluation table ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Sakana Fugu; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Sakana Fugu

80.2%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Muse Spark 1.1; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

80%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

gemini-3.6-flash via Gemini API; default sampling; single-attempt / pass@1 unless noted in Google eval methodology (July 2026). Terminus-2 harness; Terminal-Bench 2.1.

source label without registered configuration ID

78%

2.1Terminus 2 harness

Score input

provider-reported

Gemini 3.6 Flash product page with evaluation table ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Ornith-1.0-397B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Ornith-1.0-397B

77.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Artificial Analysis Terminal-Bench v2.1 independent evaluation (as listed on AA / BenchLM).

source label without registered configuration ID

77.5%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Gemini 3.6 Flash (Artificial Analysis) ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant GPT-5.3 Codex; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

GPT-5.3-Codex

77.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Gemini 3.5 Flash; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

76.2%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant GPT-5.4; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

75.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Claude Opus 4.8; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

74.6%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Terminus-2; as published in Gemini 3.6 Flash evaluation table for Gemini 3.1 Pro.

source label without registered configuration ID

73.8%

2.1Terminus 2 harness

Public reference · not scored

provider-reported

Gemini 3.6 Flash product page with evaluation table ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Qwen3.7 Plus; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

70.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Laguna S 2.1; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Laguna S 2.1

70.2%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Qwen3.7 Max; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

69.7%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Claude Opus 4.7 (Adaptive); bulk export does not retain a complete upstream harness configuration.

claude-opus-4-7-max

69.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Composer 2.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Composer 2.5

69.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant MiMo-V2.5-Pro; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

68.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant DeepSeek V4 Pro (Max); bulk export does not retain a complete upstream harness configuration.

deepseek-v4-pro-max

67.9%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Kimi K2.6; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

66.7%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant MiniMax M3; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

66%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant MiMo-V2.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

65.8%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Claude Opus 4.6; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

65.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Qwen 3.6 Max (preview); bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Qwen 3.6 Max (preview)

65.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Ornith-1.0-35B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Ornith-1.0-35B

64.2%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Inkling; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Inkling

63.8%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant GLM-5.1; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

63.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant DeepSeek V4 Pro (High); bulk export does not retain a complete upstream harness configuration.

deepseek-v4-pro-high

63.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Composer 2; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Composer 2

61.7%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Qwen3.6 Plus; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Qwen3.6 Plus

61.6%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant GPT-5.4 mini; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

60%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Step 3.7 Flash; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Step 3.7 Flash

59.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Claude Opus 4.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Claude Opus 4.5

59.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Qwen3.6-27B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

59.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Claude Sonnet 4.6; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Claude Sonnet 4.6

59.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant DeepSeek V4 Pro; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

59.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Muse Spark; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Muse Spark

59%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant MiniMax M2.7; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

MiniMax M2.7

57%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant DeepSeek V4 Flash (Max); bulk export does not retain a complete upstream harness configuration.

deepseek-v4-flash-max

56.9%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant DeepSeek V4 Flash (High); bulk export does not retain a complete upstream harness configuration.

deepseek-v4-flash-high

56.6%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Nemotron 3 Ultra; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

56.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant GLM-5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

GLM-5

56.2%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Hy3 Preview; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Hy3 Preview

54.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Gemini 3.5 Flash-Lite; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

54%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

gemini-3.5-flash-lite via Gemini API; configuration as stated in the July 2026 Google launch materials. Terminal-Bench 2.1 as published in Google launch blog.

source label without registered configuration ID

54%

2.1Provider evaluation harness

Public reference · not scored

provider-reported

Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Qwen3.5 397B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

52.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Qwen3.6-35B-A3B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

51.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Kimi K2.5 (Reasoning); bulk export does not retain a complete upstream harness configuration.

kimi-k2-5-thinking

50.8%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Kimi K2.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

50.8%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Claude Sonnet 4.5; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

50%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Qwen3.5-122B-A10B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

49.4%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant DeepSeek V4 Flash; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

49.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Grok 4.20; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Grok 4.20

47.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant GPT-5.4 nano; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

GPT-5.4 nano

46.3%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant MAI-Thinking-1; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

MAI-Thinking-1

46%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Laguna M.1; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Laguna M.1

45.8%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Ornith-1.0-9B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Ornith-1.0-9B

43.1%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Qwen3.5-27B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Qwen3.5 27B

41.6%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant GLM-4.7; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

41%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Qwen3.5-35B-A3B; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Qwen3.5-35B-A3B

40.5%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Exact BenchLM registry variant Laguna XS.2; bulk export does not retain a complete upstream harness configuration.

source label without registered configuration ID

Laguna XS.2

35.7%

2026Source-native system

Source carrier · no additional scoring weight

source-checked

BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

As published in Gemini 3.5 Flash-Lite launch materials (prior Flash-Lite comparison).

source label without registered configuration ID

Gemini 3.1 Flash-Lite

31%

2.1Provider evaluation harness

Public reference · not scored

provider-reported

Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber ↗Observed 2026-07-21 · checked 2026-07-21

Exact model variant as listed on the BenchLM public terminal-bench leaderboard page (verified 2026-07-21).

source label without registered configuration ID

69.4%

2.0/2.1BenchLM aggregated public evaluation

Score input

source-checked

Terminal-Bench 2.0 Leaderboard & Scores — July 2026 ↗Observed 2026-07-20 · checked 2026-07-21

Exact model variant as listed on the BenchLM public terminal-bench leaderboard page (verified 2026-07-21).

source label without registered configuration ID

Qwen3.6 Max

65.4%

2.0/2.1BenchLM aggregated public evaluation

Score input

source-checked

Terminal-Bench 2.0 Leaderboard & Scores — July 2026 ↗Observed 2026-07-20 · checked 2026-07-21

Exact model variant as listed on the BenchLM public terminal-bench leaderboard page (verified 2026-07-21).

source label without registered configuration ID

Hy3

54.4%

2.0/2.1BenchLM aggregated public evaluation

Score input

source-checked

Terminal-Bench 2.0 Leaderboard & Scores — July 2026 ↗Observed 2026-07-20 · checked 2026-07-21

KimiCode harness; reasoning max; Terminal-Bench 2.1 as published in the Kimi K3 launch blog.

kimi-k3-max

88.3%

2.1KimiCode

Public reference · not scored

provider-reported

Kimi K3: Open Frontier Intelligence ↗Observed 2026-07-16 · checked 2026-07-21

Muse Spark 1.1 (xhigh); Artificial Analysis independent evaluation configuration as published.

muse-spark-1-1-xhigh

77.9%

2.1Artificial Analysis independent evaluation

Public reference · not scored

source-checked

Muse Spark 1.1 (xhigh) analysis ↗Observed 2026-07-16 · checked 2026-07-16

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

91.9%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

88.0%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for gpt-5-6-sol ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

88.0%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for gpt-5-6-terra ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

88%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

87.4%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

84.7%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

84.6%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for claude-fable-5 ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

84.6%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for claude-opus-4-8 ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

84.3%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

84.3%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for gpt-5-5 ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

83.3%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

82%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

81.7%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for grok-4-5 ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

81%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

80.9%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for gpt-5-6-luna ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

80.5%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for claude-sonnet-5 ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

80.4%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

80%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

78.3%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for gpt-5-4 ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

76.2%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

74.6%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

74.5%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for qwen3-7-max ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

73.8%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for gemini-3-1-pro-preview ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

70.3%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

69.7%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

68.4%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Kimi K2.7 Code

source label without registered configuration ID

Kimi K2.7 Code

67.4%

2.1Artificial Analysis independent evaluation

Public reference · not scored

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

66%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Kimi K2.6

source label without registered configuration ID

65.9%

2.1Artificial Analysis independent evaluation

Public reference · not scored

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

65.2%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for minimax-m3 ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

64.0%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for deepseek-v4-pro ↗Observed 2026-07-15 · checked 2026-07-15

MiMo-V2.5

source label without registered configuration ID

63.7%

2.1Artificial Analysis independent evaluation

Public reference · not scored

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

61.8%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for deepseek-v4-flash ↗Observed 2026-07-15 · checked 2026-07-15

Qwen3.6 27B (Reasoning)

source label without registered configuration ID

60.7%

2.1Artificial Analysis independent evaluation

Public reference · not scored

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

GPT-5.4 mini (xhigh)

gpt-5-4-mini-xhigh

59.2%

2.1Artificial Analysis independent evaluation

Public reference · not scored

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

59.1%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Claude 4.5 Sonnet (Reasoning)

source label without registered configuration ID

55.8%

2.1Artificial Analysis independent evaluation

Score input

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Nemotron 3 Ultra 550B A55B (Reasoning)

source label without registered configuration ID

53.9%

2.1Artificial Analysis independent evaluation

Score input

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Qwen3.5 397B A17B (Reasoning)

source label without registered configuration ID

51.3%

2.1Artificial Analysis independent evaluation

Score input

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

Mistral Medium 3.5

50.6%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for mistral-medium-3-5 ↗Observed 2026-07-15 · checked 2026-07-15

GLM-4.6 (Reasoning)

source label without registered configuration ID

GLM-4.6

49.4%

2.1Artificial Analysis independent evaluation

Score input

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Exact model variant as listed on the BenchLM public benchmark leaderboard page.

source label without registered configuration ID

49.1%

2.1BenchLM aggregated public evaluation

Source carrier · no additional scoring weight

source-checked

BenchLM public benchmark leaderboards ↗Observed 2026-07-15 · checked 2026-07-15

Qwen3.5 122B A10B (Reasoning)

source label without registered configuration ID

Qwen3.5 122B

47.6%

2.1Artificial Analysis independent evaluation

Score input

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

DeepSeek V3.2 (Reasoning)

source label without registered configuration ID

DeepSeek V3.2

46.8%

2.1Artificial Analysis independent evaluation

Score input

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

GLM-4.7 (Reasoning)

source label without registered configuration ID

45.3%

2.1Artificial Analysis independent evaluation

Public reference · not scored

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

DeepSeek V3.1 Terminus (Reasoning)

source label without registered configuration ID

DeepSeek V3.1 Terminus

44.9%

2.1Artificial Analysis independent evaluation

Score input

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

Grok 4.3

39.7%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for grok-4-3 ↗Observed 2026-07-15 · checked 2026-07-15

Gemma 4 26B A4B (Reasoning)

source label without registered configuration ID

Gemma 4 26B

39.0%

2.1Artificial Analysis independent evaluation

Score input

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

NVIDIA Nemotron 3 Super 120B A12B (Reasoning)

source label without registered configuration ID

Nemotron 3 Super

38.6%

2.1Artificial Analysis independent evaluation

Score input

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Claude 4 Sonnet (Reasoning)

source label without registered configuration ID

Claude Sonnet 4

36.3%

2.1Artificial Analysis independent evaluation

Score input

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

GPT-5 (high)

gpt-5-high

GPT-5

35.2%

2.1Artificial Analysis independent evaluation

Public reference · not scored

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Nova 2.0 Pro Preview (medium)

source label without registered configuration ID

Nova 2 Pro

29.6%

2.1Artificial Analysis independent evaluation

Score input

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Gemini 2.5 Pro

source label without registered configuration ID

Gemini 2.5 Pro

28.5%

2.1Artificial Analysis independent evaluation

Public reference · not scored

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Command A+

source label without registered configuration ID

Command A+

22.8%

2.1Artificial Analysis independent evaluation

Public reference · not scored

source-checked

Terminal-Bench Hard Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

Mistral Small 4

21.0%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for mistral-small-4 ↗Observed 2026-07-15 · checked 2026-07-15

Artificial Analysis public model evaluation page; exact provider variant as listed on the page.

source label without registered configuration ID

Mistral Large 3

12.0%

2.1Artificial Analysis evaluation harness

Public reference · not scored

source-checked

Artificial Analysis evaluations for mistral-large-3 ↗Observed 2026-07-15 · checked 2026-07-15

Codex; reasoning max; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID

78.4%

2.1Codex

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-11 · checked 2026-08-16

Codex; reasoning max; Terminal-Bench 2.1 verified submission.

gpt-5-6-terra-max

78.4%

2.1Codex

Public reference · not scored

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-11 · checked 2026-07-15

Codex; reasoning max; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID

75.7%

2.1Codex

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-11 · checked 2026-08-16

Codex; reasoning max; Terminal-Bench 2.1 verified submission.

gpt-5-6-luna-max

75.7%

2.1Codex

Public reference · not scored

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-11 · checked 2026-07-15

Muse Spark 1.1; provider evaluation configuration as published in the Meta Muse Spark 1.1 evaluation report (Terminal-Bench 2.0).

muse-spark-1-1-xhigh

80%

2.1Provider evaluation harness

Public reference · not scored

provider-reported

Meta AI Muse Spark 1.1 evaluation report ↗Observed 2026-07-09 · checked 2026-07-16

Cursor CLI; reasoning high; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID

79.3%

2.1Cursor CLI

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-08-16

Claude Code; reasoning high; Terminal-Bench 2.1 verified submission.

claude-opus-4-8-high

78.9%

2.1Claude Code

Public reference · not scored

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-07-15

Claude Code; reasoning high; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID

78.9%

2.1Claude Code

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-08-16

mini-SWE-agent; reasoning xhigh; Terminal-Bench 2.1 verified submission.

muse-spark-1-1-xhigh

76.2%

2.1mini-SWE-agent

Public reference · not scored

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-07-16

mini-SWE-agent; reasoning xhigh; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID

76.2%

2.1mini-SWE-agent

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-08-16

Claude Code; reasoning high; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID

74.6%

2.1Claude Code

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-08-16

Claude Code; reasoning high; Terminal-Bench 2.1 verified submission.

claude-sonnet-5-high

74.6%

2.1Claude Code

Public reference · not scored

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-07-09 · checked 2026-07-15

LongCat-2.0 (in-house harness)

longcat-2-0-default

LongCat-2.0

70.8%

2.1LongCat unified in-house harness

Score input

provider-reported

LongCat-2.0 model card ↗Observed 2026-06-30 · checked 2026-08-15

Claude Code; reasoning xhigh; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID

83.8%

2.1Claude Code

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-06-07 · checked 2026-08-16

Claude Code; reasoning xhigh; Terminal-Bench 2.1 verified submission.

claude-fable-5-xhigh

83.8%

2.1Claude Code

Public reference · not scored

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-06-07 · checked 2026-07-15

Terminus 2; reasoning high; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID

80.5%

2.1Terminus 2

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-06-05 · checked 2026-08-16

Terminus 2; reasoning high; Terminal-Bench 2.1 verified submission.

claude-fable-5-high

80.4%

2.1Terminus 2

Public reference · not scored

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-06-05 · checked 2026-07-15

Terminus-2 agent system; provider-published Terminal-Bench 2.1 evaluation.

gpt-5-5-high

78.2%

2.1Terminus 2

Public reference · not scored

provider-reported

Gemini 3.5 Flash model card ↗Observed 2026-05-19 · checked 2026-07-15

Terminus-2 agent system; provider-published Terminal-Bench 2.1 evaluation.

gemini-3-5-flash-high

76.2%

2.1Terminus 2

Public reference · not scored

provider-reported

Gemini 3.5 Flash model card ↗Observed 2026-05-19 · checked 2026-07-15

Terminus-2 agent system; provider-published Terminal-Bench 2.1 evaluation.

source label without registered configuration ID

70.3%

2.1Terminus 2

Source carrier · no additional scoring weight

provider-reported

Gemini 3.5 Flash model card ↗Observed 2026-05-19 · checked 2026-07-15

Codex; reasoning xhigh; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID

83.2%

2.1Codex

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-08-16

Codex; reasoning xhigh; Terminal-Bench 2.1 verified submission.

gpt-5-5-xhigh

83.1%

2.1Codex

Public reference · not scored

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-07-15

Terminus 2; reasoning xhigh; Terminal-Bench 2.1 verified submission.

gpt-5-5-xhigh

78%

2.1Terminus 2

Public reference · not scored

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-07-15

Terminus 2; reasoning xhigh; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID

78.0%

2.1Terminus 2

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-08-16

Opus 4.7; reasoning=max; agent=Claude Code

claude-opus-4-7-max

68.9%

2.1Claude Code

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-08-16

Opus 4.7; reasoning=max; agent=Terminus 2

claude-opus-4-7-max

66.1%

2.1Terminus 2

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-08-16

Gemini CLI; reasoning high; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID

65.8%

2.1Gemini CLI

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-07-15

Terminus 2; reasoning high; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID

65.6%

2.1Terminus 2

Score input

official-leaderboard

Terminal-Bench 2.1 leaderboard ↗Observed 2026-05-01 · checked 2026-07-15

Claude Code; reasoning max; Terminal-Bench 2.1 verified submission.

source label without registered configuration ID