HLE leaderboard
36 models ranked, highest score first.
| # | Model | Company | Score |
|---|---|---|---|
| 1 | Claude Fable 5 | Anthropic | 53.3% |
| 2 | GLM-5 | Zhipu AI | 50.4% |
| 3 | Claude Opus 4.8 | Anthropic | 45.7% |
| 4 | GPT-5.3 Codex | OpenAI | 39.9% |
| 5 | Qwen3.7 Max | Alibaba | 38.1% |
| 6 | MiniMax-M3 | MiniMax | 37.1% |
| 7 | GPT-5.2 | OpenAI | 35.4% |
| 8 | Qwen3.7 Plus | Alibaba | 33.4% |
| 9 | Claude Opus 4.5 | Anthropic | 28.4% |
| 10 | GPT-5.1 | OpenAI | 26.5% |
| 11 | Gemini 3.5 Flash | 23.1% | |
| 12 | DeepSeek V3.2 | DeepSeek | 22.2% |
| 13 | Step 3.7 Flash | StepFun | 19.9% |
| 14 | Ring-2.6-1T | InclusionAI | 18.3% |
| 15 | HyperNova 60B 2605 | Multiverse Computing | 15.1% |
| 16 | Gemma 4 12B | 14.8% | |
| 17 | Gemini 2.5 Flash | 11.1% | |
| 18 | Claude Sonnet 4.6 | Anthropic | 10.8% |
| 19 | North Mini Code | Cohere | 9.9% |
| 20 | LFM2.5-8B-A1B | Liquid AI | 6.9% |
| 21 | MiniCPM5-1B | OpenBMB | 6.5% |
| 22 | JT-35B-Flash | China Mobile | 6.1% |
| 23 | GPT-5 | OpenAI | 5.4% |
| 24 | MiniCPM-V 4.6 1.3B | OpenBMB | 4.9% |
| 25 | Gemini 2.0 Flash | 4.7% | |
| 26 | Claude 2.1 | Anthropic | 4.2% |
| 27 | Mistral Large 3 | Mistral | 4.1% |
| 28 | Claude 3 Haiku | Anthropic | 3.9% |
| 29 | Gemini 1.5 Pro | 3.9% | |
| 30 | Grok-2 | xAI | 3.8% |
| 31 | Claude 3 Sonnet | Anthropic | 3.8% |
| 32 | Claude 3.5 Sonnet | Anthropic | 3.7% |
| 33 | DeepSeek-V3 | DeepSeek | 3.6% |
| 34 | Mistral Large | Mistral | 3.4% |
| 35 | GPT-4 Turbo | OpenAI | 3.3% |
| 36 | Claude 3 Opus | Anthropic | 3.1% |