HyperAI
Command Palette
Search for a command to run...
CritPt
一项基准，用于测试 LLM 在研究级物理推理任务上的表现，包含 71 个复合研究挑战。
- 模型
- 265 当前，600 总共
- 中位数
- 0.3%
- 范围
- 0.0%–35.7%
| 排名 | 比较 | 模型 | 评分 |
|---|---|---|---|
| 51 |  Qwen3.8 Max (0902) Alibaba | 17.7% | |
| 52 |  GPT-6 Luna (Xhigh) OpenAI | 17.4% | |
| 53 | 16.9% | ||
| 54 | GPT-5.3 Codex (Xhigh) OpenAI | 16.9% | |
| 55 |  Agnes 2.5 Pro Beta Sapiens AI | 15.7% | |
| 56 |  GLM-5.3-Flash Z AI | 15.4% | |
| 57 | GPT-6 Luna (High) OpenAI | 15.4% | |
| 58 | Agnes 2.5 Pro Theta Sapiens AI | 15.1% | |
| 59 | GPT-6.1 Terra (Low) OpenAI | 15.1% | |
| 60 | 14.9% | ||
| 61 | GLM-5.3 (Low) Z AI | 14.6% | |
| 62 |  JT-4.1 Flash 236B A21B (Reasoning) China Mobile | 14.0% | |
| 63 |  Grok 4.7 (Low) SpaceXAI | 13.4% | |
| 64 |  Beam 1.0 Reflection | 13.1% | |
| 65 |  Claude Haiku 5.5 (Medium) Anthropic | 12.9% | |
| 66 | 12.3% | ||
| 67 |  MiMo-V2.6-Flash Xiaomi | 12.0% | |
| 68 | 11.4% | ||
| 69 |  Mistral Large 4 (Medium) Mistral | 11.1% | |
| 70 | Qwen3.8-Flash-Next Alibaba | 11.1% | |
| 71 |  DeepSeek V4 Flash Vision (Max) DeepSeek | 10.9% | |
| 72 | GPT-6 Luna (Medium) OpenAI | 10.6% | |
| 73 |  Nemotron 3.5 Super VL NVIDIA | 10.6% | |
| 74 |  Kimi K2.7 Code Kimi | 10.0% | |
| 75 | Mistral Large 4 (High) Mistral | 10.0% | |
| 76 |  Quasar 438B (Max, Based on GLM-5.2) Multiverse Computing | 9.4% | |
| 77 | Claude Haiku 5.5 (Low) Anthropic | 9.1% | |
| 78 | Qwen3.7 Plus Alibaba | 9.1% | |
| 79 |  A.X-K2 SK Telecom | 8.9% | |
| 80 | 8.6% | ||
| 81 |  Inkling Small Thinking Machines | 8.3% | |
| 82 |  Motif 3 Motif Technologies | 6.6% | |
| 83 |  Solar Pro 4 Upstage | 6.3% | |
| 84 | Solar Open2 250B Upstage | 5.7% | |
| 85 | Inkling (Xhigh) Thinking Machines | 5.4% | |
| 86 | Qwen3.8 27B (Xhigh) Alibaba | 5.4% | |
| 87 | 4.9% | ||
| 88 |  Hy3 Tencent | 4.9% | |
| 89 |  Apodex 1.1 Apodex | 4.6% | |
| 90 |  K2 Horizon 375B A23B Institute of Foundation Models | 4.6% | |
| 91 | 4.3% | ||
| 92 |  Gemini 3.8 Flash (Low) Google | 4.0% | |
| 93 |  MiniMax-M3 MiniMax | 3.7% | |
| 94 |  Ring-2.6-1T InclusionAI | 3.7% | |
| 95 | 3.1% | ||
| 96 | GPT-6 Luna (Low) OpenAI | 2.6% | |
| 97 | Ling-3.0-flash-Fin InclusionAI | 2.6% | |
| 98 |  LongCat 2.0 LongCat | 2.6% | |
| 99 | 2.6% | ||
| 100 | K2 Horizon MoVA 36B A4B Institute of Foundation Models | 2.3% |
