CritPt
Unpublished research-level physics problems.
Agent Index weight
0%
shown for context only
Models tested
162
from the benchmark’s leaderboard and reports
Results
| # | Model | Score | Normalised | Harness | Cost/task | Date |
|---|---|---|---|---|---|---|
| 1 | Gemini 4 Carbon Supported Google | 36.0% | source | — | — | |
| 2 | Claude Fable 5.2 Supported Anthropic | 35.7% | source | — | — | |
| 3 | GPT-5.6 Sol Supported OpenAI | 32.3% | source | — | — | |
| 4 | Claude Opus 5.5 Supported Anthropic | 31.7% | source | — | — | |
| 5 | GPT-6.1 Sol Estimated OpenAI | 31.7% | source | — | — | |
| 6 | GPT-6 Astra Supported OpenAI | 31.7% | source | — | — | |
| 7 | Claude Sonnet 5.5 Estimated Anthropic | 31.4% | source | — | — | |
| 8 | Claude Fable 5.1 Supported Anthropic | 31.1% | source | — | — | |
| 9 | GPT-6 Sol Estimated OpenAI | 30.9% | source | — | — | |
| 10 | GPT-5.5 Pro OpenAI | 30.6% | source | — | — | |
| 11 | GPT-5.4 Pro OpenAI | 30.0% | source | — | — | |
| 12 | GPT-5.6 Terra Estimated OpenAI | 30.0% | source | — | — | |
| 13 | Claude Opus 5 Supported Anthropic | 29.1% | source | — | — | |
| 14 | Claude Fable 5 Supported Anthropic | 28.6% | source | — | — | |
| 15 | Muse Wonder Estimated Meta | 28.3% | source | — | — | |
| 16 | GPT-6.1 Terra Estimated OpenAI | 28.0% | source | — | — | |
| 17 | Gemini 4 Argon Supported Google | 27.1% | source | — | — | |
| 18 | GPT-5.5 Supported OpenAI | 27.1% | source | — | — | |
| 19 | MiMo-V2.6-Pro Xiaomi | 26.6% | source | — | — | |
| 20 | Muse Spark 1.3 Estimated Meta | 26.0% | source | — | — | |
| 21 | Gemini 3 Deep Think Google | 25.7% | source | — | — | |
| 22 | Kimi K3.1 Pro Moonshot AI | 24.6% | source | — | — | |
| 23 | DeepSeek V4.1 Flash 1017 Estimated DeepSeek | 24.0% | source | — | — | |
| 24 | GPT-5.4 Supported OpenAI | 23.4% | source | — | — | |
| 25 | Kimi K3 Supported Moonshot AI | 23.4% | source | — | — | |
| 26 | Claude Haiku 5.5 Estimated Anthropic | 22.6% | source | — | — | |
| 27 | Claude Opus 4.8 Supported Anthropic | 20.9% | source | — | — | |
| 28 | GLM-5.2 Supported Z.ai (Zhipu) | 20.9% | source | — | — | |
| 29 | Step 5 Preview stepfun | 20.9% | source | — | — | |
| 30 | GPT-5.6 Luna Estimated OpenAI | 20.6% | source | — | — | |
| 31 | Qwen3.8 Max Preview Estimated Alibaba (Qwen) | 20.0% | source | — | — | |
| 32 | Grok 4.6 Supported xAI | 19.7% | source | — | — | |
| 33 | GPT-6 Luna OpenAI | 19.4% | source | — | — | |
| 34 | GLM-5.3 Supported Z.ai (Zhipu) | 19.1% | source | — | — | |
| 35 | DeepSeek V4 Pro 0813 DeepSeek | 18.6% | source | — | — | |
| 36 | Gemini 3.8 Flash Supported Google | 18.3% | source | — | — | |
| 37 | Grok 4.7 Estimated xAI | 18.0% | source | — | — | |
| 38 | Gemini 3.1 Pro Preview Supported Google | 17.7% | source | — | — | |
| 39 | Muse Spark 1.2 Estimated Meta | 17.7% | source | — | — | |
| 40 | Qwen3.8 Max 0902 Alibaba (Qwen) | 17.7% | source | — | — | |
| 41 | Claude Sonnet 5 Supported Anthropic | 16.9% | source | — | — | |
| 42 | GPT-5.3 Codex Supported OpenAI | 16.9% | source | — | — | |
| 43 | DeepSeek V4 Flash 0731 DeepSeek | 16.6% | source | — | — | |
| 44 | GLM-5.3-Flash Supported Z.ai (Zhipu) | 15.4% | source | — | — | |
| 45 | Grok 4.5 Estimated xAI | 15.4% | source | — | — | |
| 46 | Muse Spark 1.1 Estimated Meta | 15.1% | source | — | — | |
| 47 | Muse Spark Open Meta | 14.9% | source | — | — | |
| 48 | DeepSeek V4.1 Flash Estimated DeepSeek | 14.3% | source | — | — | |
| 49 | Gemini 3.7 Flash Estimated Google | 14.3% | source | — | — | |
| 50 | Qwen3.7 Max Alibaba (Qwen) | 13.4% | source | — | — | |
| 51 | Beam 1.0 Reflection | 13.1% | source | — | — | |
| 52 | Gemini 3.5 Flash Supported Google | 13.1% | source | — | — | |
| 53 | DeepSeek V4 Pro Estimated DeepSeek | 12.9% | source | — | — | |
| 54 | GPT-5 Supported OpenAI | 12.6% | source | — | — | |
| 55 | Claude Opus 4.7 Supported Anthropic | 12.0% | source | — | — | |
| 56 | MiMo-V2.6-Flash Xiaomi | 12.0% | source | — | — | |
| 57 | Muse Spark Meta | 11.3% | source | — | — | |
| 58 | Mistral Large 4 Mistral | 11.1% | source | — | — | |
| 59 | Gemini 3.6 Flash Estimated Google | 10.6% | source | — | — | |
| 60 | GPT-5.4 mini OpenAI | 10.0% | source | — | — | |
| 61 | Kimi K2.7 Code Estimated Moonshot AI | 10.0% | source | — | — | |
| 62 | GPT-5.4 nano OpenAI | 9.3% | source | — | — | |
| 63 | Grok Build 0.1 xAI | 9.1% | source | — | — | |
| 64 | Qwen3.7 Plus Alibaba (Qwen) | 9.1% | source | — | — | |
| 65 | Inkling Small thinkingmachines | 8.3% | source | — | — | |
| 66 | Grok 4.3 Beta xAI | 8.0% | source | — | — | |
| 67 | Kimi K2.6 Estimated Moonshot AI | 8.0% | source | — | — | |
| 68 | DeepSeek V4 Flash DeepSeek | 7.1% | source | — | — | |
| 69 | Gemini 3 Pro Preview Supported Google | 6.9% | source | — | — | |
| 70 | Inkling Estimated thinkingmachines | 5.4% | source | — | — | |
| 71 | Qwen3.8 27B Alibaba (Qwen) | 5.4% | source | — | — | |
| 72 | GPT-5.1 Supported OpenAI | 4.9% | source | — | — | |
| 73 | GLM-5.1 Estimated Z.ai (Zhipu) | 4.6% | source | — | — | |
| 74 | MiMo-V2.5-Pro Xiaomi | 4.0% | source | — | — | |
| 75 | MiMo-V2.5 Xiaomi | 3.7% | source | — | — | |
| 76 | MiniMax-M3 Estimated MiniMax | 3.7% | source | — | — | |
| 77 | Ring-2.6-1T Ant Group | 3.7% | source | — | — | |
| 78 | Claude Sonnet 4.6 Supported Anthropic | 3.1% | source | — | — | |
| 79 | Kimi K2.5 Supported Moonshot AI | 3.1% | source | — | — | |
| 80 | Nemotron 3 Super NVIDIA | 3.1% | source | — | — | |
| 81 | Nemotron 3 Ultra NVIDIA | 3.1% | source | — | — | |
| 82 | DeepSeek V3.2 Supported DeepSeek | 2.9% | source | — | — | |
| 83 | Qwen3.6 Plus Estimated Alibaba (Qwen) | 2.9% | source | — | — | |
| 84 | Muse Glimmer Meta | 2.6% | source | — | — | |
| 85 | Step 3.7 Flash stepfun | 2.3% | source | — | — | |
| 86 | Gemini 2.5 Pro Supported Google | 2.0% | source | — | — | |
| 87 | DeepSeek V3.1 Terminus DeepSeek | 1.7% | source | — | — | |
| 88 | GLM-4.7 Estimated Z.ai (Zhipu) | 1.7% | source | — | — | |
| 89 | Gemma 4 31B IT Google | 1.4% | source | — | — | |
| 90 | GPT OSS 20B OpenAI | 1.4% | source | — | — | |
| 91 | Claude Sonnet 4.5 (latest) Supported Anthropic | 1.1% | source | — | — | |
| 92 | Gemini 3.1 Flash Lite Google | 1.1% | source | — | — | |
| 93 | GLM-4.6 Z.ai (Zhipu) | 1.1% | source | — | — | |
| 94 | GPT OSS 120B Supported OpenAI | 1.1% | source | — | — | |
| 95 | o3 Supported OpenAI | 1.1% | source | — | — | |
| 96 | DeepSeek-R1 Estimated DeepSeek | 1.1% | source | — | — | |
| 97 | Gemini 2.5 Flash Estimated Google | 1.1% | source | — | — | |
| 98 | Qwen3.5 122B-A10B Alibaba (Qwen) | 0.9% | source | — | — | |
| 99 | Qwen3.5 397B-A17B Alibaba (Qwen) | 0.9% | source | — | — | |
| 100 | Qwen3.6 27B Alibaba (Qwen) | 0.9% | source | — | — | |
| 101 | Trinity Large Thinking arcee-ai | 0.9% | source | — | — | |
| 102 | Mercury 2 inception | 0.8% | source | — | — | |
| 103 | o4-mini Estimated OpenAI | 0.6% | source | — | — | |
| 104 | MiniMax-M2.7 MiniMax | 0.6% | source | — | — | |
| 105 | Qwen3.5 35B-A3B Alibaba (Qwen) | 0.6% | source | — | — | |
| 106 | Qwen3.5 9B Alibaba (Qwen) | 0.6% | source | — | — | |
| 107 | Claude-Opus-4 Supported Anthropic | 0.3% | source | — | — | |
| 108 | Claude Sonnet 4 Supported Anthropic | 0.3% | source | — | — | |
| 109 | Command A+ Cohere | 0.3% | source | — | — | |
| 110 | Magistral Medium 1.2 Mistral | 0.3% | source | — | — | |
| 111 | Magistral Small 1.2 Mistral | 0.3% | source | — | — | |
| 112 | o3-mini Estimated OpenAI | 0.3% | source | — | — | |
| 113 | Qwen3-30B-A3B-Thinking (Jul 2025) Alibaba (Qwen) | 0.3% | source | — | — | |
| 114 | Qwen3 32B Alibaba (Qwen) | 0.3% | source | — | — | |
| 115 | Qwen3.6 35B-A3B Alibaba (Qwen) | 0.3% | source | — | — | |
| 116 | Amazon Nova Micro Amazon | 0.0% | source | — | — | |
| 117 | Claude 3.5 Haiku Anthropic | 0.0% | source | — | — | |
| 118 | Claude Haiku 4.5 Estimated Anthropic | 0.0% | source | — | — | |
| 119 | Cogito v2.1 Deep Cogito | 0.0% | source | — | — | |
| 120 | Cohere Command A Cohere | 0.0% | source | — | — | |
| 121 | DeepSeek V3 0324 DeepSeek | 0.0% | source | — | — | |
| 122 | DeepSeek-V3 DeepSeek | 0.0% | source | — | — | |
| 123 | Devstral Small 2 Mistral | 0.0% | source | — | — | |
| 124 | Gemini 3.5 Flash Lite Google | 0.0% | source | — | — | |
| 125 | Gemma 3 12B Google | 0.0% | source | — | — | |
| 126 | Gemma 3 270M Google | 0.0% | source | — | — | |
| 127 | Gemma 3 27B Google | 0.0% | source | — | — | |
| 128 | Gemma 4 26B A4B Google | 0.0% | source | — | — | |
| 129 | GPT-4.1 mini OpenAI | 0.0% | source | — | — | |
| 130 | GPT-4.1 nano OpenAI | 0.0% | source | — | — | |
| 131 | GPT-4o (2024-08-06) Supported OpenAI | 0.0% | source | — | — | |
| 132 | GPT Chat Latest OpenAI | 0.0% | source | — | — | |
| 133 | GPT-5 Mini Estimated OpenAI | 0.0% | source | — | — | |
| 134 | granite-4.0-350m Unknown | 0.0% | source | — | — | |
| 135 | granite-4.0-micro Unknown | 0.0% | source | — | — | |
| 136 | Granite 4.1 30B IBM | 0.0% | source | — | — | |
| 137 | Llama 3.1-405B Meta | 0.0% | source | — | — | |
| 138 | Llama 3.1-8B Meta | 0.0% | source | — | — | |
| 139 | Llama 3.2 11B Meta | 0.0% | source | — | — | |
| 140 | Llama 3.3 70B Meta | 0.0% | source | — | — | |
| 141 | Llama 4 Maverick Meta | 0.0% | source | — | — | |
| 142 | Llama 4 Scout Meta | 0.0% | source | — | — | |
| 143 | Mercury 2.5 inception | 0.0% | source | — | — | |
| 144 | Mimo-V2-Flash Xiaomi | 0.0% | source | — | — | |
| 145 | Mistral Large 3 Mistral | 0.0% | source | — | — | |
| 146 | Mistral Medium 3.1 Mistral | 0.0% | source | — | — | |
| 147 | Mistral Medium 3.5 Mistral | 0.0% | source | — | — | |
| 148 | Mistral Small 3.2 Mistral | 0.0% | source | — | — | |
| 149 | Mistral Small 3.1 Mistral | 0.0% | source | — | — | |
| 150 | Nvidia Nemotron Nano 9B V2 NVIDIA | 0.0% | source | — | — | |
| 151 | Nova 2.0 Pro Preview Amazon | 0.0% | source | — | — | |
| 152 | Phi 4 Microsoft | 0.0% | source | — | — | |
| 153 | Phi 4 Mini Microsoft | 0.0% | source | — | — | |
| 154 | Qwen3 14B Alibaba (Qwen) | 0.0% | source | — | — | |
| 155 | Qwen3 235B A22B Thinking 2507 Alibaba (Qwen) | 0.0% | source | — | — | |
| 156 | Qwen3.5 0.8B Alibaba (Qwen) | 0.0% | source | — | — | |
| 157 | Qwen3.5 2B Alibaba (Qwen) | 0.0% | source | — | — | |
| 158 | Qwen3.5 4B Alibaba (Qwen) | 0.0% | source | — | — | |
| 159 | Qwen3 8B Alibaba (Qwen) | 0.0% | source | — | — | |
| 160 | Qwen3 Coder Next Alibaba (Qwen) | 0.0% | source | — | — | |
| 161 | Reka Flash 3 rekaai | 0.0% | source | — | — | |
| 162 | solar-pro3 upstage | 0.0% | source | — | — |