Humanity's Last Exam (AA)
Artificial Analysis' own HLE run. Shown for reference.
- Published by
- Artificial Analysis
- Category
- Reasoning
- Index weight
- Reference only
- Models
- 524
- Data as of
- 9 Sept 2026
Source: Artificial Analysis (artificialanalysis.ai). Data taken from the public leaderboard.
- 1Claude Fable 5.159.1%
- 2Claude Fable 5.158.7%
- 3Claude Fable 5.155.9%
- 4Claude Fable 555.5%
- 5Claude Opus 554.9%
- 6GPT-6 Astra54.7%
- 7GPT-6 Astra54.6%
- 8Claude Opus 554.4%
- 9Claude Fable 5.153.8%
- 10GPT-6 Astra53.1%
- 11Claude Opus 552.8%
- 12GPT-6 Astra52.7%
- 13Claude Opus 551.3%
- 14GPT-5.6 Sol49.5%
- 15GPT-6 Astra49.2%
524 of 524
| # | |||||
|---|---|---|---|---|---|
| 1 | 59.1% | Claude Fable 5.1 (max with fallback) | 70.0 | – | |
| 2 | 58.7% | – | 70.3 | – | |
| 3 | 55.9% | – | 69.7 | – | |
| 4 | 55.5% | Claude Fable 5 (with fallback) | 70.4 | – | |
| 5 | 54.9% | Claude Opus 5 (max) | 70.0 | – | |
| 6 | 54.7% | GPT-6 Astra (max) | 69.9 | – | |
| 7 | 54.6% | – | 68.6 | – | |
| 8 | 54.4% | – | 66.3 | – | |
| 9 | 53.8% | – | 68.0 | – | |
| 10 | 53.1% | – | 69.5 | – | |
| 11 | 52.8% | – | 69.3 | – | |
| 12 | 52.7% | – | 67.4 | – | |
| 13 | 51.3% | – | 63.1 | – | |
| 14 | 49.5% | GPT-5.6 Sol (max) | 65.3 | – | |
| 15 | 49.2% | – | 65.8 | – | |
| 16 | 48.9% | – | 65.9 | – | |
| 17 | 48.7% | Muse Spark 1.3 (max) | – | – | |
| 18 | 48.7% | Claude Opus 4.8 (max) | 64.7 | – | |
| 19 | 47.9% | Gemini 3.7 Flash (high) | 61.7 | – | |
| 20 | 47.8% | Gemini 3.8 Flash (high) | 62.4 | – | |
| 21 | 47.5% | – | 63.9 | – | |
| 22 | 47.3% | – | 67.5 | – | |
| 23 | 47.0% | – | 63.9 | – | |
| 24 | 46.9% | Kimi K3 (max) | 66.5 | – | |
| 25 | 46.2% | Muse Spark 1.1 (xhigh) | 64.9 | – | |
| 26 | 46.0% | – | 68.0 | – | |
| 27 | 45.8% | GPT-5.5 (xhigh) | 66.6 | – | |
| 28 | 45.5% | Muse Spark 1.2 (xhigh) | 62.6 | – | |
| 29 | 45.0% | – | 66.3 | – | |
| 30 | 44.1% | – | 62.3 | – | |
| 31 | 43.7% | GPT-5.4 (xhigh) | 63.4 | – | |
| 32 | 43.4% | – | 61.7 | – | |
| 33 | 43.0% | – | 63.9 | – | |
| 34 | 42.9% | Grok 4.6 (high) | 63.6 | – | |
| 35 | 42.9% | GPT-5.6 Terra (max) | 62.3 | – | |
| 36 | 42.7% | Grok 4.5 (high) | 63.4 | – | |
| 37 | 42.7% | Gemini 3.5 Flash | 60.7 | – | |
| 38 | 42.5% | GPT-5.3 Codex (xhigh) | 62.4 | – | |
| 39 | 42.5% | – | – | – | |
| 40 | 42.4% | – | 64.9 | – | |
| 41 | 42.3% | Claude Opus 4.7 (max) | 65.5 | – | |
| 42 | 42.3% | GLM-5.3 (max) | – | – | |
| 43 | 42.2% | – | 66.0 | – | |
| 44 | 42.1% | – | 63.8 | – | |
| 45 | 42.1% | – | 63.5 | – | |
| 46 | 41.9% | – | 64.2 | – | |
| 47 | 41.3% | – | 63.9 | – | |
| 48 | 41.3% | Claude Sonnet 5 (max) | 61.2 | – | |
| 49 | 41.1% | GLM-5.2 (max) | 57.4 | – | |
| 50 | 41.0% | DeepSeek V4 Pro 0813 (max) | 59.7 | – | |
| 51 | 40.8% | Gemini 3.6 Flash | 60.7 | – | |
| 52 | 40.7% | – | 65.4 | – | |
| 53 | 40.5% | – | 58.9 | – | |
| 54 | Motif 3 (Beta)Motif Technologies | 40.4% | – | – | – |
| 55 | 39.9% | – | 63.1 | – | |
| 56 | 39.9% | – | 61.0 | – | |
| 57 | 39.7% | Gemini 3 Pro Preview (high) | 61.6 | – | |
| 58 | 39.5% | GPT-5.6 Luna (max) | 56.6 | – | |
| 59 | 39.4% | – | 64.6 | – | |
| 60 | 39.0% | – | 57.3 | – |