Our data / Logic & Reasoning
Lumina Logic & Reasoning Index
By LuminaIndex designed by LuminaBench using attributed, publisher-run benchmarks. Lumina did not rerun those evaluations.LMCA, DTBench and ACCoRD.
Lumina Logic & Reasoning Index
Complete results · 3/3 · 20 of 154 models · best variant per model
Benchmark
Metric
Lumina Logic & Reasoning Index vs Cost per Task
Original average USD per LMCA/DTBench response as documented in the source data README; absent costs stay unavailable. No time or token measurements inferred.
15 of 154 families
Results table
| Rank | Model / configuration | Lumina Score | Coverage |
|---|---|---|---|
| 1 | Claude Opus 5.5 Max | 1,132 | 3/3 · Observed |
| 2 | Claude Sonnet 5.5 Max | 1,128 | 3/3 · Observed |
| 3 | Claude Fable 5.1 Max | 1,118 | 3/3 · Observed |
| 4 | Claude Fable 5 Max | 1,117 | 3/3 · Observed |
| 5 | GPT-6 Astra Max | 1,116 | 3/3 · Observed |
| 6 | Claude Opus 5 Max | 1,116 | 3/3 · Observed |
| 7 | GPT-6.1 Sol Not specified | 1,112 | 3/3 · Observed |
| 8 | GPT-6 Sol Not specified | 1,110 | 3/3 · Observed |
| 9 | Gemini 4 Argon High | 1,105 | 3/3 · Observed |
| 10 | GPT-5.6 Sol Pro Not specified | 1,103 | 3/3 · Observed |
| 11 | GPT-5.5 Not specified | 1,097 | 3/3 · Observed |
| 12 | Claude Opus 4.8 Not specified | 1,096 | 3/3 · Observed |
| 13 | GPT-5.6 Sol Max | 1,096 | 3/3 · Observed |
| 14 | Kimi K3.1 Pro Not specified | 1,094 | 3/3 · Observed |
| 15 | Muse Wonder Not specified | 1,094 | 3/3 · Observed |
| 16 | Claude Opus 4.6 Not specified | 1,093 | 3/3 · Observed |
| 17 | Claude Haiku 5.5 Max | 1,093 | 3/3 · Observed |
| 18 | Claude Opus 4.7 Not specified | 1,091 | 3/3 · Observed |
| 19 | GPT-5.4 Not specified | 1,091 | 3/3 · Observed |
| 20 | GPT-5.5 Pro Not specified | 1,090 | 3/3 · Observed |
| 21 | Muse Spark 1.3 Not specified | 1,089 | 3/3 · Observed |
| 22 | Gemini 3.8 Flash High | 1,088 | 3/3 · Observed |
| 23 | Gemini 3.7 Flash High | 1,087 | 3/3 · Observed |
| 24 | GPT-5.6 Terra Max | 1,086 | 3/3 · Observed |
| 25 | Muse Spark 1.1 Not specified | 1,085 | 3/3 · Observed |
| 26 | Muse Spark 1.2 Not specified | 1,082 | 3/3 · Observed |
| 27 | Grok 4.7 Not specified | 1,080 | 3/3 · Observed |
| 28 | Grok 4.6 Not specified | 1,077 | 3/3 · Observed |
| 29 | Claude Sonnet 5 Max | 1,075 | 3/3 · Observed |
| 30 | Kimi K3 Not specified | 1,073 | 3/3 · Observed |
| 31 | Gemini 3.1 Pro High | 1,073 | 3/3 · Observed |
| 32 | GLM-5.3 Not specified | 1,073 | 3/3 · Observed |
| 33 | GPT-5.2 Not specified | 1,072 | 3/3 · Observed |
| 34 | Claude Opus 4.5 Not specified | 1,071 | 3/3 · Observed |
| 35 | Grok 4.5 Not specified | 1,071 | 3/3 · Observed |
| 36 | Gemini 3.6 Flash Not specified | 1,067 | 3/3 · Observed |
| 37 | Qwen3.8 2.4T A95B Not specified | 1,067 | 3/3 · Observed |
| 38 | Claude Sonnet 4.6 Not specified | 1,066 | 3/3 · Observed |
| 39 | Qwen3.8-Max Not specified | 1,066 | 3/3 · Observed |
| 40 | DeepSeek V4 Pro 0813 Not specified | 1,065 | 3/3 · Observed |
| 41 | Gemini 3.5 Flash Not specified | 1,064 | 3/3 · Observed |
| 42 | GPT-6 Luna Not specified | 1,061 | 3/3 · Observed |
| 43 | GLM-5.2 Not specified | 1,059 | 3/3 · Observed |
| 44 | GPT-5.1 Not specified | 1,059 | 3/3 · Observed |
| 45 | DeepSeek V4.1 Flash Not specified | 1,059 | 3/3 · Observed |
| 46 | GPT-5.6 Luna Max | 1,057 | 3/3 · Observed |
| 47 | GPT-5 Not specified | 1,057 | 3/3 · Observed |
| 48 | DeepSeek-V4-Pro Not specified | 1,055 | 3/3 · Observed |
| 49 | Qwen3.8-27B Not specified | 1,055 | 3/3 · Observed |
| 50 | Grok 4.20 Multi-Agent Not specified | 1,054 | 3/3 · Observed |
| 51 | Qwen3.7-Max Not specified | 1,054 | 3/3 · Observed |
| 52 | Gemini 3 Flash Not specified | 1,052 | 3/3 · Observed |
| 53 | Qwen3.6-Max (preview) Not specified | 1,052 | 3/3 · Observed |
| 54 | Kimi K2.6 Not specified | 1,050 | 3/3 · Observed |
| 55 | o3-pro Not specified | 1,050 | 3/3 · Observed |
| 56 | Grok 4.20 Not specified | 1,047 | 3/3 · Observed |
| 57 | Inkling Not specified | 1,046 | 3/3 · Observed |
| 58 | o3 Not specified | 1,045 | 3/3 · Observed |
| 59 | DeepSeek V4 Flash 0731 Not specified | 1,043 | 3/3 · Observed |
| 60 | Grok 4.3 Not specified | 1,042 | 3/3 · Observed |
| 61 | Hy3 Not specified | 1,041 | 3/3 · Observed |
| 62 | Claude Sonnet 4.5 Not specified | 1,040 | 3/3 · Observed |
| 63 | Muse Glimmer 30B Not specified | 1,040 | 3/3 · Observed |
| 64 | GPT-5.4 mini Not specified | 1,039 | 3/3 · Observed |
| 65 | Nemotron 3 Ultra Not specified | 1,038 | 3/3 · Observed |
| 66 | Mistral Large 4 Not specified | 1,036 | 3/3 · Observed |
| 67 | Gemini 3.5 Flash-Lite Not specified | 1,035 | 3/3 · Observed |
| 68 | GPT-5.2 Chat Not specified | 1,033 | 3/3 · Observed |
| 69 | GPT-5 mini Not specified | 1,031 | 3/3 · Observed |
| 70 | Claude Opus 4 Not specified | 1,029 | 3/3 · Observed |
| 71 | GPT-5.4 nano Not specified | 1,028 | 3/3 · Observed |
| 72 | Gemma 4 31B Not specified | 1,027 | 3/3 · Observed |
| 73 | Qwen3-Max (thinking) Not specified | 1,027 | 3/3 · Observed |
| 74 | Qwen3.7-Plus Not specified | 1,025 | 3/3 · Observed |
| 75 | Qwen3.5 397B-A17B Not specified | 1,023 | 3/3 · Observed |
| 76 | Claude Opus 4.1 Not specified | 1,022 | 3/3 · Observed |
| 77 | Gemini 2.5 Pro Not specified | 1,021 | 3/3 · Observed |
| 78 | DeepSeek V4 Flash Not specified | 1,021 | 3/3 · Observed |
| 79 | DeepSeek V3.2 Not specified | 1,020 | 3/3 · Observed |
| 80 | MiniMax M3 Not specified | 1,019 | 3/3 · Observed |
| 81 | GPT-5.3 Chat Not specified | 1,019 | 3/3 · Observed |
| 82 | DeepSeek-V3.2-Exp Not specified | 1,019 | 3/3 · Observed |
| 83 | Qwen3.5 27B Not specified | 1,017 | 3/3 · Observed |
| 84 | Gemini 3.1 Flash-Lite Not specified | 1,016 | 3/3 · Observed |
| 85 | Qwen3.5-122B-A10B Not specified | 1,013 | 3/3 · Observed |
| 86 | Qwen3.6-Plus Not specified | 1,013 | 3/3 · Observed |
| 87 | qwen3.5 Plus Not specified | 1,011 | 3/3 · Observed |
| 88 | MiMo-V2.5-Pro Not specified | 1,010 | 3/3 · Observed |
| 89 | Qwen3.6 27B Not specified | 1,009 | 3/3 · Observed |
| 90 | Claude Haiku 4.5 Not specified | 1,009 | 3/3 · Observed |
| 91 | Qwen Plus Not specified | 1,008 | 3/3 · Observed |
| 92 | Qwen3-Max Not specified | 1,008 | 3/3 · Observed |
| 93 | Qwen3.6-Flash Not specified | 1,006 | 3/3 · Observed |
| 94 | o4-mini Not specified | 1,003 | 3/3 · Observed |
| 95 | Qwen3 235B-A22B Thinking (2507) Not specified | 1,001 | 3/3 · Observed |
| 96 | GPT-Oss 120b Not specified | 1,000 | 3/3 · Observed |
| 97 | DeepSeek-V3.1-Terminus Not specified | 999 | 3/3 · Observed |
| 98 | Qwen3.5 35B-A3B Not specified | 998 | 3/3 · Observed |
| 99 | Qwen3.5-Flash Not specified | 998 | 3/3 · Observed |
| 100 | Claude Sonnet 4 Not specified | 996 | 3/3 · Observed |
| 101 | DeepSeek Chat V3.1 Not specified | 995 | 3/3 · Observed |
| 102 | o1 Not specified | 993 | 3/3 · Observed |
| 103 | Qwen3 235B-A22B (2507) Not specified | 992 | 3/3 · Observed |
| 104 | Gemma 4 26B-A4B Not specified | 990 | 3/3 · Observed |
| 105 | qwen3.6 35b a3b Not specified | 987 | 3/3 · Observed |
| 106 | Qwen3-Next-80B-A3B-Thinking Not specified | 986 | 3/3 · Observed |
| 107 | GPT-4.1 Not specified | 977 | 3/3 · Observed |
| 108 | Gemini 2.5 Flash Not specified | 976 | 3/3 · Observed |
| 109 | Qwen3 235B-A22B Not specified | 973 | 3/3 · Observed |
| 110 | Mistral Medium 3.5 Not specified | 972 | 3/3 · Observed |
| 111 | Nemotron 3.5 Lightning Not specified | 971 | 3/3 · Observed |
| 112 | qwen3 30b a3b Thinking 2507 Not specified | 970 | 3/3 · Observed |
| 113 | qwen3.5 9B Not specified | 970 | 3/3 · Observed |
| 114 | Qwen3-Next-80B-A3B-Instruct Not specified | 970 | 3/3 · Observed |
| 115 | gpt-oss-20b Not specified | 968 | 3/3 · Observed |
| 116 | o3-mini Not specified | 968 | 3/3 · Observed |
| 117 | Qwen3 30B-A3B Instruct (2507) Not specified | 954 | 3/3 · Observed |
| 118 | GPT-4.1-Mini Not specified | 952 | 3/3 · Observed |
| 119 | Qwen3-32B Not specified | 951 | 3/3 · Observed |
| 120 | qwen3 14b Not specified | 948 | 3/3 · Observed |
| 121 | Mistral Large 3 Not specified | 947 | 3/3 · Observed |
| 122 | Mistral Small (Mar '26) Not specified | 947 | 3/3 · Observed |
| 123 | GPT-5 nano Not specified | 946 | 3/3 · Observed |
| 124 | Llama 4 Maverick Not specified | 944 | 3/3 · Observed |
| 125 | GPT-4o Not specified | 941 | 3/3 · Observed |
| 126 | mistral medium 3.1 Not specified | 940 | 3/3 · Observed |
| 127 | DeepSeek Chat V3 (0324) Not specified | 937 | 3/3 · Observed |
| 128 | Llama 3.3 70B Not specified | 936 | 3/3 · Observed |
| 129 | Qwen3-30B-A3B Not specified | 934 | 3/3 · Observed |
| 130 | Llama 4 Scout Not specified | 931 | 3/3 · Observed |
| 131 | DeepSeek Chat Not specified | 931 | 3/3 · Observed |
| 132 | Qwen2.5 72B Not specified | 930 | 3/3 · Observed |
| 133 | Llama 3.1 70B Not specified | 929 | 3/3 · Observed |
| 134 | Gemini 2.5 Flash-Lite Not specified | 923 | 3/3 · Observed |
| 135 | GPT-4 Turbo Not specified | 922 | 3/3 · Observed |
| 136 | Claude 3 Opus Not specified | 919 | 3/3 · Observed |
| 137 | qwen3 8b Not specified | 917 | 3/3 · Observed |
| 138 | Command A Not specified | 915 | 3/3 · Observed |
| 139 | Gemma 3 27B Not specified | 910 | 3/3 · Observed |
| 140 | Ministral 14B Not specified | 904 | 3/3 · Observed |
| 141 | GPT-4o-mini Not specified | 898 | 3/3 · Observed |
| 142 | Ministral 8B (Dec '25) Not specified | 895 | 3/3 · Observed |
| 143 | Gemma 3n E4B Not specified | 893 | 3/3 · Observed |
| 144 | Gemma 3 12B Not specified | 890 | 3/3 · Observed |
| 145 | Llama 3.1-8B Not specified | 887 | 3/3 · Observed |
| 146 | GPT-4.1 nano Not specified | 876 | 3/3 · Observed |
| 147 | Qwen2.5-7B Not specified | 875 | 3/3 · Observed |
| 148 | GPT-3.5 Turbo Not specified | 870 | 3/3 · Observed |
| 149 | Command R+ (Aug '24) Not specified | 869 | 3/3 · Observed |
| 150 | Gemma 3 4B Not specified | 868 | 3/3 · Observed |
| 151 | Claude 3 Haiku Not specified | 863 | 3/3 · Observed |
| 152 | Ministral 3B Not specified | 860 | 3/3 · Observed |
| 153 | Command R (Aug '24) Not specified | 856 | 3/3 · Observed |
| 154 | Command R7B Not specified | 845 | 3/3 · Observed |
| — | Claude Sonnet 5 xHigh | — | 2/3 · Source only |
| — | Claude 2.0 Not specified | — | 1/3 · Source only |
| — | Claude Sonnet 5.5 Low | — | 2/3 · Source only |
| — | Claude Fable 5 High | — | 2/3 · Source only |
| — | Claude Sonnet 5.5 High | — | 2/3 · Source only |
| — | Command R (Mar '24) Not specified | — | 1/3 · Source only |
| — | Claude Haiku 5.5 off | — | 2/3 · Source only |
| — | GPT-5.6 Sol Medium | — | 2/3 · Source only |
| — | Claude Sonnet 5.5 xHigh | — | 2/3 · Source only |
| — | Claude Fable 5.1 Low | — | 2/3 · Source only |
| — | Claude Fable 5.1 High | — | 2/3 · Source only |
| — | Reka Core (Jul '24) Not specified | — | 1/3 · Source only |
| — | Gemini 1.0 Pro Not specified | — | 1/3 · Source only |
| — | Llama 2 13B Not specified | — | 1/3 · Source only |
| — | Claude Sonnet 5 High | — | 2/3 · Source only |
| — | Claude Opus 5 Medium | — | 2/3 · Source only |
| — | GPT-5.6 Sol xHigh | — | 2/3 · Source only |
| — | GPT-6 Astra Medium | — | 2/3 · Source only |
| — | Claude Opus 5.5 High | — | 2/3 · Source only |
| — | Aya Expanse 32B Not specified | — | 1/3 · Source only |
| — | GPT-5.6 Luna High | — | 2/3 · Source only |
| — | GPT-5.6 Terra Medium | — | 2/3 · Source only |
| — | Claude Haiku 5.5 High | — | 2/3 · Source only |
| — | Gemini 1.5 Flash-8B Not specified | — | 1/3 · Source only |
| — | Reka Core (May '24) Not specified | — | 1/3 · Source only |
| — | GPT-5.6 Terra Low | — | 2/3 · Source only |
| — | Gemini 2.0 Flash-Lite Not specified | — | 1/3 · Source only |
| — | Gemini 1.5 Flash (002) Not specified | — | 1/3 · Source only |
| — | Gemini 1.5 Flash (001) Not specified | — | 1/3 · Source only |
| — | Mistral Small 3.2 24B Not specified | — | 1/3 · Source only |
| — | Mistral Medium (Aug '25) Not specified | — | 1/3 · Source only |
| — | Command Not specified | — | 1/3 · Source only |
| — | Mistral Large (Nov '24) Not specified | — | 1/3 · Source only |
| — | Gemma 2 27B Not specified | — | 2/3 · Source only |
| — | Mistral Small 3.2 (Jun '25) Not specified | — | 1/3 · Source only |
| — | GPT-6 Astra High | — | 2/3 · Source only |
| — | Mistral Medium (Dec '23) Not specified | — | 1/3 · Source only |
| — | Claude Sonnet 5 Low | — | 2/3 · Source only |
| — | GPT-5.6 Luna Low | — | 2/3 · Source only |
| — | Llama 2 70B Not specified | — | 1/3 · Source only |
| — | Mistral Large (Feb '24) Not specified | — | 1/3 · Source only |
| — | Gemini 3.7 Flash Low | — | 2/3 · Source only |
| — | GPT-5.6 Luna Medium | — | 2/3 · Source only |
| — | Claude Opus 5 Low | — | 2/3 · Source only |
| — | Mistral Small (Sep '24) Not specified | — | 1/3 · Source only |
| — | Claude Opus 5.5 Low | — | 2/3 · Source only |
| — | Claude Fable 5 Low | — | 2/3 · Source only |
| — | GPT-6 Astra xHigh | — | 2/3 · Source only |
| — | Mistral Small 3.1 24B Not specified | — | 1/3 · Source only |
| — | GPT-5.6 Sol off | — | 2/3 · Source only |
| — | Mixtral 8x22B Not specified | — | 1/3 · Source only |
| — | Claude Fable 5.1 Medium | — | 2/3 · Source only |
| — | Claude Sonnet 5.5 Medium | — | 2/3 · Source only |
| — | Claude Opus 5 xHigh | — | 2/3 · Source only |
| — | Claude 3.5 Sonnet (Jun '24) Not specified | — | 1/3 · Source only |
| — | Mistral Medium 3 (May '25) Not specified | — | 1/3 · Source only |
| — | Claude Opus 5 High | — | 2/3 · Source only |
| — | Claude Opus 5.5 xHigh | — | 2/3 · Source only |
| — | GPT-5.6 Sol High | — | 2/3 · Source only |
| — | Command R+ (Apr '24) Not specified | — | 1/3 · Source only |
| — | Magistral Small (Sep '25) Not specified | — | 1/3 · Source only |
| — | Claude 3 Sonnet Not specified | — | 1/3 · Source only |
| — | Reka Flash (Sep '24) Not specified | — | 1/3 · Source only |
| — | Reka Core (Sep '24) Not specified | — | 1/3 · Source only |
| — | Aya Expanse 8B Not specified | — | 1/3 · Source only |
| — | GPT-4 Not specified | — | 1/3 · Source only |
| — | Grok 4.1 fast Not specified | — | 1/3 · Source only |
| — | GPT-5.6 Luna off | — | 2/3 · Source only |
| — | GPT-4 Not specified | — | 1/3 · Source only |
| — | Mistral Small (Dec '23) Not specified | — | 1/3 · Source only |
| — | Llama 3 70B Not specified | — | 1/3 · Source only |
| — | Claude Haiku 5.5 Medium | — | 2/3 · Source only |
| — | Claude Fable 5.1 xHigh | — | 2/3 · Source only |
| — | Gemini 4 Argon Medium | — | 2/3 · Source only |
| — | Reka Flash (Jul '24) Not specified | — | 1/3 · Source only |
| — | Mistral NeMo Not specified | — | 1/3 · Source only |
| — | Claude Fable 5 Medium | — | 2/3 · Source only |
| — | GPT-5.6 Sol Low | — | 2/3 · Source only |
| — | Gemini 2.0 Flash Not specified | — | 1/3 · Source only |
| — | Llama 3-8B Not specified | — | 1/3 · Source only |
| — | Llama 3.1 405B Not specified | — | 1/3 · Source only |
| — | Reka Core (Apr '24) Not specified | — | 1/3 · Source only |
| — | Reka Flash Preview (Jun '24) Not specified | — | 1/3 · Source only |
| — | Gemini 3.7 Flash Medium | — | 2/3 · Source only |
| — | Grok 2 Not specified | — | 1/3 · Source only |
| — | GPT-5.6 Luna xHigh | — | 2/3 · Source only |
| — | Claude 3.5 Haiku Not specified | — | 1/3 · Source only |
| — | Claude Opus 5.5 Medium | — | 2/3 · Source only |
| — | Claude Fable 5 xHigh | — | 2/3 · Source only |
| — | Mistral Small 3.1 (Mar '25) Not specified | — | 1/3 · Source only |
| — | Reka Flash (Feb '24) Not specified | — | 1/3 · Source only |
| — | Mistral Large 2 (Jul '24) Not specified | — | 1/3 · Source only |
| — | Gemini 3.1 Pro Medium | — | 2/3 · Source only |
| — | GPT-5.6 Terra xHigh | — | 2/3 · Source only |
| — | GPT-6 Astra Low | — | 2/3 · Source only |
| — | Mistral Small (Feb '24) Not specified | — | 1/3 · Source only |
| — | Gemini 1.5 Pro (002) Not specified | — | 1/3 · Source only |
| — | Mistral Tiny Not specified | — | 1/3 · Source only |
| — | Claude Haiku 5.5 Low | — | 2/3 · Source only |
| — | Claude Haiku 5.5 xHigh | — | 2/3 · Source only |
| — | GPT-5.6 Terra High | — | 2/3 · Source only |
| — | Gemini 3.8 Flash Low | — | 2/3 · Source only |
| — | GPT-4 Not specified | — | 1/3 · Source only |
| — | Grok 4 Fast Not specified | — | 1/3 · Source only |
| — | Gemini 1.5 Pro (001) Not specified | — | 1/3 · Source only |
| — | Claude Sonnet 5 Medium | — | 2/3 · Source only |
| — | Command (nightly) Not specified | — | 1/3 · Source only |
| — | Claude Instant 1.2 Not specified | — | 1/3 · Source only |
| — | Claude 2.1 Not specified | — | 1/3 · Source only |
| — | Ministral 8B (Oct '24) Not specified | — | 1/3 · Source only |
| — | Gemini 3.1 Pro Low | — | 2/3 · Source only |
| — | Claude 3.5 Sonnet (Oct '24) Not specified | — | 1/3 · Source only |
| — | Code Llama 34B Not specified | — | 1/3 · Source only |
| — | Gemini 4 Argon Low | — | 2/3 · Source only |
| — | GPT-5.6 Terra off | — | 2/3 · Source only |
| — | Gemini 3.8 Flash Medium | — | 2/3 · Source only |
Method & coverage
Equal thirds across compatible components of the same retained Conceptual Reasoning publisher row. LMCA measures human-rating correlation, DTBench uses original decision accuracy, and ACCoRD uses normalized constraint violation with its already oriented conversion. The parent CRI aggregate is excluded. These tasks share a research programme; three scores do not mean three independent evaluators. Unspecified effort stays unspecified, linked only by the exact shared submission, never by a wildcard. Partial/assisted coverage is retained source-only. This does not replace the existing Lumina Reasoning Index. Publisher effort-study rows omit ACCoRD away from the canonical tier; no top-tier result is copied to another setting. Base-profile links do not verify unpublished API endpoints.
154 complete measured configurations; no missing-component estimates. Other observations remain source-only on their benchmark pages. Full-precision converted scores are weighted before rounding once for display. Costs, times, tokens and tool measurements always describe the selected benchmark workload.
Method: logic-reasoning-index-observed-v2. Sources have recorded publication eligibility.