Humanity's Last Exam — General tools-enabled variant, many sources; diff pop than humanitys_last_exam_with_tools.
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | Opus 5 | Anthropic | 64.7 | 1 | 2026-07-24 |
| 2 | Claude Mythos 5 | Anthropic | 64.7 | 2 | 2026-06-12 |
| 3 | Claude Fable 5 | Anthropic | 64.5 | 3 | 2026-08-12 |
| 4 | Claude Opus 4.6 | Anthropic | 62.7 | 12 | 2026-07-04 |
| 5 | GPT-5.4 Pro | OpenAI | 58.7 | 2 | 2026-06-21 |
| 6 | GPT-5.6 Sol | OpenAI | 58 | 2 | 2026-08-12 |
| 7 | Claude Opus 4.8 | Anthropic | 57.9 | 7 | 2026-08-12 |
| 8 | Qwen3.8 Max Preview | Alibaba | 56.2 | 1 | 2026-08-12 |
| 9 | GLM-5.2 Full Open Source | Z.ai | 54.7 | 4 | 2026-07-16 |
| 10 | Claude Opus 4.7 | Anthropic | 54.7 | 3 | 2026-06-21 |
| 11 | Kimi K2.6 1T A32B | Moonshot | 54 | 1 | 2026-06-06 |
| 12 | Kimi K2.6 | Moonshot | 54 | 5 | 2026-07-16 |
| 13 | Qwen3.7 Max | Alibaba | 53.5 | 3 | 2026-08-12 |
| 14 | GLM-5.1 | Z.ai | 52.3 | 6 | 2026-07-13 |
| 15 | GPT-5.5 | OpenAI | 52.2 | 6 | 2026-08-12 |
| 16 | GPT-5.4 | OpenAI | 52.1 | 9 | 2026-07-04 |
| 17 | Kimi K2.5 | Moonshot | 51.8 | 6 | 2026-07-16 |
| 18 | Gemini 3.1 Pro | 51.6 | 16 | 2026-07-16 | |
| 19 | Qwen3.6 Plus | Alibaba | 50.6 | 2 | 2026-05-18 |
| 20 | GLM-5 | Z.ai | 50.4 | 3 | 2026-05-18 |
| 21 | GLM-5.1-744B-A40B | Z.ai | 50.4 | 1 | 2026-06-06 |
| 22 | GPT-5.6 Luna | OpenAI | 48.9 | 2 | 2026-07-30 |
| 23 | Qwen 3.5 397B 17B | Alibaba | 48.3 | 1 | 2026-06-06 |
| 24 | Qwen3.5 397B A17B | Alibaba | 48.3 | 2 | 2026-07-30 |
| 25 | DeepSeek-V4-Pro | DeepSeek | 48.2 | 9 | 2026-07-16 |
| 26 | Inkling-Small | Thinking Machines | 47.8 | 2 | 2026-07-30 |
| 27 | Claude Sonnet 4.6 | Anthropic | 46.8 | 1 | 2026-06-30 |
| 28 | Claudes | Anthropic | 46.8 | 1 | 2026-08-12 |
| 29 | Inkling | Thinking Machines | 46.6 | 3 | 2026-07-30 |
| 30 | Gemini 3 Pro | 45.8 | 3 | 2026-06-15 | |
| 31 | GPT-5.2 | OpenAI | 45.5 | 2 | 2026-06-15 |
| 32 | DeepSeek-V4-Flash | DeepSeek | 45.1 | 4 | 2026-07-30 |
| 33 | Kimi K2 (Reasoning) | Moonshot | 44.9 | 2 | 2026-06-06 |
| 34 | Claude Opus 4.5 | Anthropic | 43.4 | 2 | 2026-06-15 |
| 35 | GLM-4.7 | Z.ai | 42.8 | 2 | 2026-05-18 |
| 36 | GPT-5.1 | OpenAI | 42.7 | 1 | 2026-05-18 |
| 37 | Gemini 3.5 Flash Lite | 42.5 | 2 | 2026-07-30 | |
| 38 | DeepSeek-V3.2 | DeepSeek | 40.8 | 6 | 2026-06-15 |
| 39 | MiniMax M2.7 | MiniMax | 40.3 | 2 | 2026-07-30 |
| 40 | MiMo V2.5 Pro | Xiaomi | 40 | 2 | 2026-07-30 |
| 41 | nvidia-nemotron-3-ultra-550b-a55b | NVIDIA | 37.4 | 1 | 2026-08-24 |
| 42 | Nemotron 3 Ultra 550B A55B | NVIDIA | 37.4 | 1 | 2026-06-06 |
| 43 | GPT-5 | OpenAI | 35.2 | 2 | 2026-06-06 |
| 44 | Claude Sonnet 4.5 | Anthropic | 32 | 2 | 2026-06-06 |
| 45 | MiniMax M2 | MiniMax | 31.8 | 1 | 2026-06-06 |
| 46 | GLM-4.6 | Z.ai | 30.4 | 2 | 2026-06-06 |
| 47 | Gemini 2.5 Pro | 28.4 | 1 | 2026-06-06 | |
| 48 | Gemma 4 31B | 26.5 | 26 | 2026-07-11 | |
| 49 | Claude Sonnet 4 | Anthropic | 20.3 | 1 | 2026-06-06 |
| 50 | Gemma 4 26B A4B | 17.2 | 22 | 2026-07-16 | |
| 51 | Gemma 4 12B | 5.2 | 1 | 2026-07-16 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.