LMCA
Language-model capability assessment across domains.
Agent Index weight
0%
shown for context only
Models tested
140
from the benchmark’s leaderboard and reports
Results
| # | Model | Score | Normalised | Harness | Cost/task | Date |
|---|---|---|---|---|---|---|
| 1 | Claude Fable 5.2 Supported Anthropic | 70.6% | source | — | — | |
| 2 | Claude Sonnet 5.5 Estimated Anthropic | 68.3% | source | — | — | |
| 3 | Claude Opus 5.5 Supported Anthropic | 68.2% | source | — | — | |
| 4 | Claude Fable 5.1 Supported Anthropic | 65.5% | source | — | — | |
| 5 | Claude Opus 5 Supported Anthropic | 64.5% | source | — | — | |
| 6 | GPT-6 Astra Supported OpenAI | 64.4% | source | — | — | |
| 7 | Gemini 4 Argon Supported Google | 63.7% | source | — | — | |
| 8 | GPT-6.1 Sol Estimated OpenAI | 62.8% | source | — | — | |
| 9 | Claude Fable 5 Supported Anthropic | 61.2% | source | — | — | |
| 10 | GPT-5.6 Sol Supported OpenAI | 59.2% | source | — | — | |
| 11 | GPT-6 Sol Estimated OpenAI | 59.1% | source | — | — | |
| 12 | Kimi K3.1 Pro Moonshot AI | 57.6% | source | — | — | |
| 13 | Claude Opus 4.8 Supported Anthropic | 57.5% | source | — | — | |
| 14 | Muse Wonder Estimated Meta | 56.8% | source | — | — | |
| 15 | Claude Opus 4.6 Supported Anthropic | 55.8% | source | — | — | |
| 16 | GLM-5.3 Supported Z.ai (Zhipu) | 55.5% | source | — | — | |
| 17 | GPT-5.6 Terra Estimated OpenAI | 54.9% | source | — | — | |
| 18 | GPT-5.5 Supported OpenAI | 54.3% | source | — | — | |
| 19 | GPT-5.5 Pro OpenAI | 53.9% | source | — | — | |
| 20 | Muse Spark 1.3 Estimated Meta | 53.9% | source | — | — | |
| 21 | Gemini 3.1 Pro Preview Supported Google | 53.8% | source | — | — | |
| 22 | Gemini 3.8 Flash Supported Google | 52.9% | source | — | — | |
| 23 | Kimi K3 Supported Moonshot AI | 52.7% | source | — | — | |
| 24 | Claude Opus 4.7 Supported Anthropic | 52.2% | source | — | — | |
| 25 | GPT-5.4 Supported OpenAI | 52.0% | source | — | — | |
| 26 | Claude Haiku 5.5 Estimated Anthropic | 51.1% | source | — | — | |
| 27 | Gemini 3.7 Flash Estimated Google | 50.4% | source | — | — | |
| 28 | Claude Sonnet 5 Supported Anthropic | 50.0% | source | — | — | |
| 29 | Muse Spark 1.1 Estimated Meta | 49.9% | source | — | — | |
| 30 | Grok 4.7 Estimated xAI | 49.4% | source | — | — | |
| 31 | GPT-5.6 Luna Estimated OpenAI | 48.5% | source | — | — | |
| 32 | Grok 4.6 Supported xAI | 48.5% | source | — | — | |
| 33 | Muse Spark 1.2 Estimated Meta | 48.4% | source | — | — | |
| 34 | Gemini 3.5 Flash Supported Google | 47.1% | source | — | — | |
| 35 | DeepSeek V4.1 Flash Estimated DeepSeek | 47.0% | source | — | — | |
| 36 | Claude Sonnet 4.6 Supported Anthropic | 46.5% | source | — | — | |
| 37 | Qwen3.8 Max Preview Estimated Alibaba (Qwen) | 46.2% | source | — | — | |
| 38 | GLM-5.2 Supported Z.ai (Zhipu) | 45.8% | source | — | — | |
| 39 | DeepSeek V4 Pro 0813 DeepSeek | 45.5% | source | — | — | |
| 40 | Grok 4.5 Estimated xAI | 45.2% | source | — | — | |
| 41 | Gemini 3.6 Flash Estimated Google | 44.9% | source | — | — | |
| 42 | GPT-6 Luna OpenAI | 44.5% | source | — | — | |
| 43 | Claude Opus 4.5 (latest) Supported Anthropic | 44.5% | source | — | — | |
| 44 | Qwen3.7 Max Alibaba (Qwen) | 44.0% | source | — | — | |
| 45 | GPT-5.2 Supported OpenAI | 43.9% | source | — | — | |
| 46 | GPT-5.1 Supported OpenAI | 43.9% | source | — | — | |
| 47 | Gemini 3 Flash Preview Supported Google | 43.1% | source | — | — | |
| 48 | Qwen3.6 Max Preview Estimated Alibaba (Qwen) | 42.5% | source | — | — | |
| 49 | DeepSeek V4 Flash 0731 DeepSeek | 41.7% | source | — | — | |
| 50 | Qwen3.8 27B Alibaba (Qwen) | 41.4% | source | — | — | |
| 51 | DeepSeek V4 Pro Estimated DeepSeek | 41.2% | source | — | — | |
| 52 | GPT-5.4 mini OpenAI | 40.8% | source | — | — | |
| 53 | GPT-5 Supported OpenAI | 40.0% | source | — | — | |
| 54 | o3 Supported OpenAI | 39.7% | source | — | — | |
| 55 | Gemma 4 31B IT Google | 39.3% | source | — | — | |
| 56 | Claude Sonnet 4.5 (latest) Supported Anthropic | 38.8% | source | — | — | |
| 57 | Grok 4.20 (Non-Reasoning) Estimated xAI | 38.7% | source | — | — | |
| 58 | o3-pro OpenAI | 38.5% | source | — | — | |
| 59 | Grok 4.3 Beta xAI | 38.3% | source | — | — | |
| 60 | Qwen3.5 397B-A17B Alibaba (Qwen) | 37.9% | source | — | — | |
| 61 | Gemini 3.5 Flash Lite Google | 37.6% | source | — | — | |
| 62 | Inkling Estimated thinkingmachines | 37.6% | source | — | — | |
| 63 | Qwen3.7 Plus Alibaba (Qwen) | 37.6% | source | — | — | |
| 64 | Claude-Opus-4 Supported Anthropic | 37.4% | source | — | — | |
| 65 | Kimi K2.6 Estimated Moonshot AI | 37.3% | source | — | — | |
| 66 | Claude Opus 4.1 (latest) Supported Anthropic | 37.1% | source | — | — | |
| 67 | Nemotron 3 Ultra NVIDIA | 36.9% | source | — | — | |
| 68 | GPT-5.4 nano OpenAI | 36.9% | source | — | — | |
| 69 | Qwen3.5 Plus Estimated Alibaba (Qwen) | 36.4% | source | — | — | |
| 70 | DeepSeek V4 Flash DeepSeek | 35.9% | source | — | — | |
| 71 | Mistral Large 4 Mistral | 35.8% | source | — | — | |
| 72 | Gemini 3.1 Flash Lite Google | 35.0% | source | — | — | |
| 73 | Gemini 2.5 Pro Supported Google | 34.8% | source | — | — | |
| 74 | Qwen3.6 27B Alibaba (Qwen) | 34.5% | source | — | — | |
| 75 | GPT-5 Mini Estimated OpenAI | 34.2% | source | — | — | |
| 76 | Qwen3.5 27B Alibaba (Qwen) | 34.0% | source | — | — | |
| 77 | MiniMax-M3 Estimated MiniMax | 33.7% | source | — | — | |
| 78 | Qwen3.6 Plus Estimated Alibaba (Qwen) | 33.1% | source | — | — | |
| 79 | Qwen3.5 122B-A10B Alibaba (Qwen) | 32.2% | source | — | — | |
| 80 | Qwen3.6 Flash Alibaba (Qwen) | 31.0% | source | — | — | |
| 81 | Claude Haiku 4.5 Estimated Anthropic | 30.9% | source | — | — | |
| 82 | Qwen3.6 35B-A3B Alibaba (Qwen) | 29.7% | source | — | — | |
| 83 | Gemma 4 26B A4B Google | 29.7% | source | — | — | |
| 84 | MiMo-V2.5-Pro Xiaomi | 29.5% | source | — | — | |
| 85 | Qwen3.5 35B-A3B Alibaba (Qwen) | 29.5% | source | — | — | |
| 86 | Qwen3 235B A22B Thinking 2507 Alibaba (Qwen) | 29.3% | source | — | — | |
| 87 | Qwen3.5 Flash Alibaba (Qwen) | 29.1% | source | — | — | |
| 88 | DeepSeek V3.2 Supported DeepSeek | 29.1% | source | — | — | |
| 89 | Claude Sonnet 4 Supported Anthropic | 29.0% | source | — | — | |
| 90 | DeepSeek V3.1 Terminus DeepSeek | 28.6% | source | — | — | |
| 91 | Qwen3 Max Estimated Alibaba (Qwen) | 28.3% | source | — | — | |
| 92 | Gemini 2.5 Flash Estimated Google | 27.5% | source | — | — | |
| 93 | o4-mini Estimated OpenAI | 26.5% | source | — | — | |
| 94 | Mistral Medium 3.5 Mistral | 26.1% | source | — | — | |
| 95 | GPT-4.1 OpenAI | 25.6% | source | — | — | |
| 96 | Qwen3 235B A22B Alibaba (Qwen) | 25.0% | source | — | — | |
| 97 | Qwen3.5 9B Alibaba (Qwen) | 24.5% | source | — | — | |
| 98 | DeepSeek V3.1 DeepSeek | 24.3% | source | — | — | |
| 99 | Qwen Plus Alibaba (Qwen) | 24.0% | source | — | — | |
| 100 | Qwen3 235B A22B Instruct 2507 Alibaba (Qwen) | 23.5% | source | — | — | |
| 101 | Qwen3 30B A3B Instruct 2507 Alibaba (Qwen) | 22.4% | source | — | — | |
| 102 | o1 Estimated OpenAI | 22.3% | source | — | — | |
| 103 | GPT OSS 120B Supported OpenAI | 22.1% | source | — | — | |
| 104 | GPT-4.1 mini OpenAI | 21.1% | source | — | — | |
| 105 | Mistral Small 4 Mistral | 20.6% | source | — | — | |
| 106 | Qwen3-30B-A3B-Thinking (Jul 2025) Alibaba (Qwen) | 19.5% | source | — | — | |
| 107 | Mistral Medium 3.1 Mistral | 19.1% | source | — | — | |
| 108 | o3-mini Estimated OpenAI | 19.0% | source | — | — | |
| 109 | Qwen3 14B Alibaba (Qwen) | 18.2% | source | — | — | |
| 110 | Gemini 2.5 Flash-Lite Google | 18.1% | source | — | — | |
| 111 | Llama 3.3 70B Meta | 17.5% | source | — | — | |
| 112 | Qwen3 32B Alibaba (Qwen) | 17.3% | source | — | — | |
| 113 | GPT-4 OpenAI | 17.1% | source | — | — | |
| 114 | Claude 3 Opus Estimated Anthropic | 17.0% | source | — | — | |
| 115 | Mistral Large 3 Mistral | 16.7% | source | — | — | |
| 116 | GPT-4o (2024-08-06) Supported OpenAI | 16.6% | source | — | — | |
| 117 | Llama 4 Maverick Meta | 15.9% | source | — | — | |
| 118 | Qwen3 30B A3B Alibaba (Qwen) | 15.8% | source | — | — | |
| 119 | DeepSeek V3 0324 DeepSeek | 15.5% | source | — | — | |
| 120 | Llama 3.1-70B Meta | 14.8% | source | — | — | |
| 121 | GPT OSS 20B OpenAI | 14.5% | source | — | — | |
| 122 | Qwen2.5-72B Estimated Alibaba (Qwen) | 13.4% | source | — | — | |
| 123 | Gemma 3 27B Google | 12.3% | source | — | — | |
| 124 | Llama 4 Scout Meta | 12.0% | source | — | — | |
| 125 | GPT-4o mini OpenAI | 10.4% | source | — | — | |
| 126 | Cohere Command A Cohere | 10.3% | source | — | — | |
| 127 | GPT-4 Turbo OpenAI | 9.8% | source | — | — | |
| 128 | GPT-3.5-turbo OpenAI | 9.7% | source | — | — | |
| 129 | c4ai-command-r-08-2024 Unknown | 9.2% | source | — | — | |
| 130 | Claude-Haiku-3 Anthropic | 8.8% | source | — | — | |
| 131 | Qwen3 8B Alibaba (Qwen) | 8.8% | source | — | — | |
| 132 | GPT-5 Nano OpenAI | 7.9% | source | — | — | |
| 133 | Gemma 2 27B Google | 7.1% | source | — | — | |
| 134 | Qwen2.5-7B Alibaba (Qwen) | 6.4% | source | — | — | |
| 135 | Ministral 3B Mistral | 5.5% | source | — | — | |
| 136 | GPT-4.1 nano OpenAI | 5.5% | source | — | — | |
| 137 | Llama 3.1-8B Meta | 5.4% | source | — | — | |
| 138 | Command R+ Cohere | 5.0% | source | — | — | |
| 139 | Gemma 3 12B Google | 4.5% | source | — | — | |
| 140 | Gemma 3 4B Google | 2.8% | source | — | — |