# Scale Labs Humanity's Last Exam

> A Scale Labs public leaderboard mirrored as display-only reference data. It does not affect BenchLM rankings.

Canonical page: https://benchlm.ai/benchmarks/scale-humanitys-last-exam

- Category: [Knowledge](/knowledge)
- Last updated: October 15, 2026 snapshot

## About Humanity's Last Exam

- Year: 2026
- Tasks: 55 published rows
- Format: Published Scale leaderboard score
- Difficulty: External agent and model evaluation
- Paper: [Scale Labs leaderboard](https://labs.scale.com/leaderboard/humanitys_last_exam)

BenchLM mirrors 55 published rows from the Humanity's Last Exam public table captured on October 15, 2026 snapshot.

Humanity's Last Exam is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (55 models)

| Rank | Model | Creator | Score |
|------|-------|---------|-------|
| 1 | [GPT-6 Astra](/models/gpt-6-astra) | OpenAI | 54.80% |
| 2 | [Fable 5.1 (xhigh)\n](https://labs.scale.com/leaderboard/humanitys_last_exam) | Anthropic | 46.50% |
| 3 | [gemini-3.1-pro-preview (thinking high)](https://labs.scale.com/leaderboard/humanitys_last_exam) | Google | 46.44% |
| 4 | [Sonnet 5.5 (xhigh)](https://labs.scale.com/leaderboard/humanitys_last_exam) | Anthropic | 46.12% |
| 5 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | Google | 44.52% |
| 6 | [GPT-5.4 Pro](/models/gpt-5-4-pro) | OpenAI | 44.32% |
| 7 | [Muse Spark](/models/muse-spark) | Meta | 40.56% |
| 8 | [gemini-3-pro-preview](https://labs.scale.com/leaderboard/humanitys_last_exam) | Google | 37.52% |
| 9 | [gpt-5.4-2026-03-05 (xhigh thinking)](https://labs.scale.com/leaderboard/humanitys_last_exam) | OpenAI | 36.24% |
| 10 | [Claude Opus 4.7](/models/claude-opus-4-7) | Anthropic | 36.20% |
| 11 | [claude-opus-4-6-thinking-max](https://labs.scale.com/leaderboard/humanitys_last_exam) | Anthropic | 34.44% |
| 12 | [gpt-5-pro-2025-10-06](https://labs.scale.com/leaderboard/humanitys_last_exam) | OpenAI | 31.64% |
| 13 | [Claude Haiku 5.5](/models/claude-haiku-5-5) | Anthropic | 30.12% |
| 14 | [GPT-5.2](/models/gpt-5-2) | OpenAI | 27.80% |
| 15 | [gpt-5-2025-08-07](https://labs.scale.com/leaderboard/humanitys_last_exam) | OpenAI | 25.32% |
| 16 | [Claude Opus 4.5 Thinking](/models/claude-opus-4-5-thinking) | Anthropic | 25.20% |
| 17 | [Kimi K2.5](/models/kimi-k2-5) | Moonshot AI | 24.37% |
| 18 | [gpt-5.1-thinking](https://labs.scale.com/leaderboard/humanitys_last_exam) | OpenAI | 23.68% |
| 19 | [gemini-2.5-pro-preview-06-05](https://labs.scale.com/leaderboard/humanitys_last_exam) | Google | 21.64% |
| 20 | [o3 (high) (April 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam) | OpenAI | 20.32% |
| 21 | [GPT-5 mini](/models/gpt-5-mini) | OpenAI | 19.44% |
| 22 | [o3 (medium) (April 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam) | OpenAI | 19.20% |
| 23 | [claude-opus-4-6 (Non-Thinking)](https://labs.scale.com/leaderboard/humanitys_last_exam) | Anthropic | 19.00% |
| 24 | [Gemini 2.5 Pro Experimental (March 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam) | Google | 18.16% |
| 25 | [o4-mini (high) (April 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam) | OpenAI | 18.08% |
| 26 | [Gemini 2.5 Pro Preview (May 06 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam) | Google | 17.80% |
| 27 | [o4-mini (medium) (April 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam) | OpenAI | 14.28% |
| 28 | [Claude Opus 4.5](/models/claude-opus-4-5) | Anthropic | 14.16% |
| 29 | [Claude Sonnet 4.5 Thinking](/models/claude-sonnet-4-5-thinking) | Anthropic | 13.72% |
| 30 | [Gemini 2.5 Flash (April 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam) | Google | 12.08% |
| 31 | [claude-opus-4-1-20250805-thinking](https://labs.scale.com/leaderboard/humanitys_last_exam) | Anthropic | 11.52% |
| 32 | [Gemini 2.5 Flash Preview (May 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam) | Google | 10.96% |
| 33 | [Claude Opus 4 (Thinking)](https://labs.scale.com/leaderboard/humanitys_last_exam) | Anthropic | 10.72% |
| 34 | [gemini-3.1-flash-lite-preview](https://labs.scale.com/leaderboard/humanitys_last_exam) | Google | 8.64% |
| 35 | [glm-4p5](https://labs.scale.com/leaderboard/humanitys_last_exam) | Z.AI | 8.32% |
| 36 | [glm-4p5-air](https://labs.scale.com/leaderboard/humanitys_last_exam) | Z.AI | 8.12% |
| 37 | [o1-pro](/models/o1-pro) | OpenAI | 8.12% |
| 38 | [Claude 3.7 Sonnet (Thinking)](https://labs.scale.com/leaderboard/humanitys_last_exam) | Anthropic | 8.04% |
| 39 | [o1 (December 2024)](https://labs.scale.com/leaderboard/humanitys_last_exam) | OpenAI | 7.96% |
| 40 | [claude-opus-4-1-20250805](https://labs.scale.com/leaderboard/humanitys_last_exam) | Anthropic | 7.92% |
| 41 | [Claude Sonnet 4 (Thinking)\n](https://labs.scale.com/leaderboard/humanitys_last_exam) | Anthropic | 7.76% |
| 42 | [Claude Sonnet 4.5](/models/claude-sonnet-4-5) | Anthropic | 7.52% |
| 43 | [gpt-5.1-instant](https://labs.scale.com/leaderboard/humanitys_last_exam) | OpenAI | 6.80% |
| 44 | [Claude Opus 4](https://labs.scale.com/leaderboard/humanitys_last_exam) | Anthropic | 6.68% |
| 45 | [Gemini 2.0 Flash Thinking (January 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam) | Google | 6.56% |
| 46 | [Llama 4 Maverick](/models/llama-4-maverick) | Meta | 5.68% |
| 47 | [Claude Sonnet 4](https://labs.scale.com/leaderboard/humanitys_last_exam) | Anthropic | 5.52% |
| 48 | [GPT 4.5 Preview](https://labs.scale.com/leaderboard/humanitys_last_exam) | OpenAI | 5.44% |
| 49 | [GPT-4.1](/models/gpt-4-1) | OpenAI | 5.40% |
| 50 | [Gemini-1.5-Pro-002](https://labs.scale.com/leaderboard/humanitys_last_exam) | Google | 4.60% |
| 51 | [Mistral Medium 3](/models/mistral-medium-3) | Mistral | 4.52% |
| 52 | [Nova Pro](/models/nova-pro) | Amazon | 4.40% |
| 53 | [Claude 3.5 Sonnet (October 2024)](https://labs.scale.com/leaderboard/humanitys_last_exam) | Anthropic | 4.08% |
| 54 | [Nova Lite](https://labs.scale.com/leaderboard/humanitys_last_exam) | Amazon | 3.64% |
| 55 | [GPT-4o (November 2024)](https://labs.scale.com/leaderboard/humanitys_last_exam) | OpenAI | 2.72% |

## FAQ

### What does Humanity's Last Exam measure?

A Scale Labs public leaderboard mirrored as display-only reference data. It does not affect BenchLM rankings.

### Which model leads the published Humanity's Last Exam snapshot?

GPT-6 Astra currently leads the published Humanity's Last Exam snapshot with a score of 54.80%.

### How many models are evaluated on Humanity's Last Exam?

The October 15, 2026 snapshot contains 55 AI models.

### Does Humanity's Last Exam affect BenchLM's overall score?

Not directly. Humanity's Last Exam is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.