Instruction Following · Benchmark profile

Instruction Following Benchmark

Reference2025Active17 models

Data verified 1 Aug 2026 · Methodology 2.3.0

Benchmark score on Instruction Following Benchmark

Leaderboard · 17 models

MAI-Thinking-1 leads at 85%, followed by Inkling-Small (82.2%) and Nemotron 3 Ultra (81.7%).

Display onlyInstruction Following2025

#1Closed

MAI-Thinking-1

Microsoft

mai-thinking-1

85%

Overall — · Context 256K tokens

#2Open

Inkling-Small

Thinking Machines Lab

thinkingmachines/inkling-small

82.2%

Overall — · Context 524K tokens

#3Open

Nemotron 3 Ultra

NVIDIA

nemotron-3-ultra-550b

81.7%

Overall #71 · Context 262K tokens

ViewModelsColours

MicrosoftThinking Machines LabNVIDIAxAIAlibaba CloudGoogle

Lumina

Instruction Following Benchmark leaderboard

Accessible data for Instruction Following Benchmark leaderboard
Model	Rank	Model ID	Provider	Score	Unit
MAI-Thinking-1	#1	mai-thinking-1	Microsoft	85	percent
Inkling-Small	#2	inkling-small	Thinking Machines Lab	82.2	percent
Nemotron 3 Ultra	#3	nemotron-3-ultra	NVIDIA	81.7	percent
Grok 4.3	#4	grok-4-3	xAI	81.3	percent
Inkling	#5	inkling	Thinking Machines Lab	79.8	percent
Qwen3.7-Max	#6	qwen-3-7-max	Alibaba Cloud	79.1	percent
Qwen3.7-Plus	#7	qwen-3-7-plus	Alibaba Cloud	79.1	percent
Gemini 3.5 Flash	#8	gemini-3-5-flash	Google	76.3	percent
Qwen3.6 Plus	#9	qwen3-6-plus	Alibaba Cloud	75.8	percent
Nemotron 3 Nano Omni 30B A3B	#10	nemotron-3-nano-omni-30b-a3b	NVIDIA	74.2	percent
Hy3 Preview	#11	hy3-preview	Tencent	63.1	percent
Claude Opus 4.5	#12	claude-opus-4-5	Anthropic	58	percent
Ling 2.6 Flash	#13	ling-2-6-flash	InclusionAI	57	percent
LFM2.5-8B-A1B	#14	lfm2-5-8b-a1b	LiquidAI	56.47	percent
ZAYA1-8B	#15	zaya1-8b	Zyphra	52.56	percent
MiniCPM5-1B	#16	minicpm5-1b	OpenBMB	46.67	percent
LFM2.5-230M	#17	lfm2-5-230m	LiquidAI	38.4	percent

Benchmark score table (17 models)

One best compatible score per canonical product · higher is better

Instruction Following Benchmark leaderboard ranked by score
Rank	Model	Provider	License	Evidence use	Score
#1	MAI-Thinking-1 mai-thinking-1	Microsoft	closed	Source carrier	85%
#2	Inkling-Small thinkingmachines/inkling-small	Thinking Machines Lab	open	Source carrier	82.2%
#3	Nemotron 3 Ultra nemotron-3-ultra-550b	NVIDIA	open	Source carrier	81.7%
#4	Grok 4.3 grok-4.3	xAI	closed	Source carrier	81.3%
#5	Inkling inkling	Thinking Machines Lab	open	Source carrier	79.8%
#6	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	closed	Source carrier	79.1%
#7	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	closed	Source carrier	79.1%
#8	Gemini 3.5 Flash gemini-3.5-flash	Google	closed	Source carrier	76.3%
#9	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	Source carrier	75.8%
#10	Nemotron 3 Nano Omni 30B A3B nemotron-3-nano-omni-30b-a3b	NVIDIA	open	Source carrier	74.2%
#11	Hy3 Preview hy3-preview	Tencent	open	Source carrier	63.1%
#12	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	Source carrier	58%
#13	Ling 2.6 Flash ling-2-6-flash	InclusionAI	open	Source carrier	57%
#14	LFM2.5-8B-A1B lfm2-5-8b-a1b	LiquidAI	open	Source carrier	56.5%
#15	ZAYA1-8B zaya1-8b	Zyphra	open	Source carrier	52.6%
#16	MiniCPM5-1B minicpm5-1b	OpenBMB	open	Source carrier	46.7%
#17	LFM2.5-230M lfm2-5-230m	LiquidAI	open	Source carrier	38.4%

The top of this snapshot is led by MAI-Thinking-1 at 85%; third place is 3.3 points behind. The top-10 spread is 10.8 points.

Configuration-preserving evidencePublished source rows (49)Exact source labels, configurations, versions and dated observations. Same-snapshot canonical/configuration projections are shown once.

Every distinct dated source-published row for Instruction Following Benchmark, including exact model configuration labels
Exact published model / configuration	Canonical product	Score	Version / system	Evidence	Source / observed
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	85%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
Exact BenchLM registry variant Inkling-Small; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Inkling-Small	82.2%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
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	81.7%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
Exact BenchLM registry variant Grok 4.3; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Grok 4.3	81.3%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
Exact BenchLM registry variant Inkling; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Inkling	79.8%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
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	79.1%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
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	79.1%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
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	76.3%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
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	75.8%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
Exact BenchLM registry variant Nemotron 3 Nano Omni 30B A3B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Nemotron 3 Nano Omni 30B A3B	74.2%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
Exact BenchLM registry variant Hy3 Preview; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Hy3 Preview	63.1%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
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	58%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
Exact BenchLM registry variant Ling 2.6 Flash; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Ling 2.6 Flash	57%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
Exact BenchLM registry variant LFM2.5-8B-A1B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	LFM2.5-8B-A1B	56.5%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
Exact BenchLM registry variant ZAYA1-8B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	ZAYA1-8B	52.6%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
Exact BenchLM registry variant MiniCPM5-1B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MiniCPM5-1B	46.7%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
Exact BenchLM registry variant LFM2.5-230M; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	LFM2.5-230M	38.4%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 2026-08-01 ↗Observed 2026-08-01 · checked 2026-08-01
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	85%	2025Source-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	81.7%	2025Source-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 4.3; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Grok 4.3	81.3%	2025Source-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	79.8%	2025Source-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	79.1%	2025Source-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	79.1%	2025Source-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	76.3%	2025Source-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	75.8%	2025Source-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 Nano Omni 30B A3B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Nemotron 3 Nano Omni 30B A3B	74.2%	2025Source-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	63.1%	2025Source-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	58%	2025Source-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 Ling 2.6 Flash; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Ling 2.6 Flash	57%	2025Source-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 LFM2.5-8B-A1B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	LFM2.5-8B-A1B	56.5%	2025Source-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 ZAYA1-8B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	ZAYA1-8B	52.6%	2025Source-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 MiniCPM5-1B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MiniCPM5-1B	46.7%	2025Source-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 LFM2.5-230M; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	LFM2.5-230M	38.4%	2025Source-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	85%	2025Source-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	81.7%	2025Source-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 4.3; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Grok 4.3	81.3%	2025Source-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	79.8%	2025Source-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	79.1%	2025Source-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 Plus; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Qwen3.7-Plus	79.1%	2025Source-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	76.3%	2025Source-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	75.8%	2025Source-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 Nano Omni 30B A3B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Nemotron 3 Nano Omni 30B A3B	74.2%	2025Source-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	63.1%	2025Source-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	58%	2025Source-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 Ling 2.6 Flash; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	Ling 2.6 Flash	57%	2025Source-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 LFM2.5-8B-A1B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	LFM2.5-8B-A1B	56.5%	2025Source-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 ZAYA1-8B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	ZAYA1-8B	52.6%	2025Source-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 MiniCPM5-1B; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	MiniCPM5-1B	46.7%	2025Source-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 LFM2.5-230M; bulk export does not retain a complete upstream harness configuration. source label without registered configuration ID	LFM2.5-230M	38.4%	2025Source-native system	Source carrier · no additional scoring weight source-checked	BenchLM public datasets — 21 July 2026 ↗Observed 2026-07-21 · checked 2026-07-21

Compare top models

MAI-Thinking-1 vs Inkling-Small Inkling-Small vs Nemotron 3 Ultra Nemotron 3 Ultra vs Grok 4.3 Grok 4.3 vs Inkling

About Instruction Following Benchmark

Definition and scoring

Organisation: BenchLM registry
Category: Instruction Following
Version: 2025
Direction: higher is better
Ranking use: Reference
Contamination risk: Unknown

IFBench evaluates precise instruction-following generalization on 58 challenging, verifiable out-of-domain constraints. Unlike IFEval which tests familiar constraint types, IFBench specifically measures how well models follow novel instructions they haven't been optimized for, exposing overfitting to common instruction patterns. Every genuine source row stays tied to its exact model label, configuration, benchmark version and evaluation system. The summary chart shows one best compatible score per canonical product; Score 2.0 admits only explicitly mapped, frozen protocols and keeps incompatible configurations separate.

Open benchmark source ↗

FAQ

Quick answers

What does Instruction Following Benchmark measure?

Which model scores highest on Instruction Following Benchmark?

MAI-Thinking-1 by Microsoft currently leads with 85%.

How many models are evaluated on Instruction Following Benchmark?

17 canonical model products have a qualifying score, backed by 49 distinct dated source rows.

Instruction Following · Benchmark profile

Instruction Following Benchmark

Reference2025Active17 models

Data verified 1 Aug 2026 · Methodology 2.3.0

Benchmark score on Instruction Following Benchmark