# Scale Labs Humanity's Last Exam (Text Only)

> 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-text-only

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

## About Humanity's Last Exam (Text Only)

- Year: 2026
- Tasks: 65 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_text_only)

BenchLM mirrors 65 published rows from the Humanity's Last Exam (Text Only) public table captured on October 15, 2026 snapshot.

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

## Leaderboard (65 models)

| Rank | Model | Creator | Score |
|------|-------|---------|-------|
| 1 | [GPT-6 Astra](/models/gpt-6-astra) | OpenAI | 54.17% |
| 2 | [gemini-3.1-pro-preview (thinking high)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Google | 47.31% |
| 3 | [Fable 5.1 (xhigh)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Anthropic | 46.80% |
| 4 | [Sonnet 5.5 (xhigh)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Anthropic | 46.62% |
| 5 | [GPT-5.4 Pro](/models/gpt-5-4-pro) | OpenAI | 45.32% |
| 6 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | Google | 44.95% |
| 7 | [Muse Spark](/models/muse-spark) | Meta | 40.92% |
| 8 | [gemini-3-pro-preview](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Google | 37.72% |
| 9 | [gpt-5.4-2026-03-05 (xhigh thinking)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 36.47% |
| 10 | [claude-opus-4-6-thinking-max](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Anthropic | 36.24% |
| 11 | [gpt-5-pro-2025-10-06](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 33.32% |
| 12 | [Claude Haiku 5.5](/models/claude-haiku-5-5) | Anthropic | 30.35% |
| 13 | [GPT-5.2](/models/gpt-5-2) | OpenAI | 28.50% |
| 14 | [Claude Opus 4.5 Thinking](/models/claude-opus-4-5-thinking) | Anthropic | 26.32% |
| 15 | [gpt-5-2025-08-07](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 26.32% |
| 16 | [gpt-5.1-thinking](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 24.65% |
| 17 | [gemini-2.5-pro-preview-06-05](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Google | 22.06% |
| 18 | [o3 (high) (April 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 20.57% |
| 19 | [o3 (medium) (April 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 19.78% |
| 20 | [GPT-5 mini](/models/gpt-5-mini) | OpenAI | 19.74% |
| 21 | [claude-opus-4-6 (Non-Thinking)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Anthropic | 19.37% |
| 22 | [o4-mini (high) (April 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 18.90% |
| 23 | [Gemini 2.5 Pro Experimental (March 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Google | 18.38% |
| 24 | [Gemini 2.5 Pro Preview (May 06 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Google | 18.38% |
| 25 | [GPT-OSS 120B](/models/gpt-oss-120b) | OpenAI | 15.48% |
| 26 | [Qwen3-235B-A22B-Thinking-2507](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Alibaba | 15.43% |
| 27 | [o4-mini (medium) (April 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 14.53% |
| 28 | [Claude Sonnet 4.5 Thinking](/models/claude-sonnet-4-5-thinking) | Anthropic | 14.09% |
| 29 | [DeepSeek-R1-0528](/models/deepseek-r1-0528) | DeepSeek | 14.04% |
| 30 | [Claude Opus 4.5](/models/claude-opus-4-5) | Anthropic | 13.90% |
| 31 | [o3 mini (high)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 13.37% |
| 32 | [DeepSeek V3.1](/models/deepseek-v3-1) | DeepSeek | 12.88% |
| 33 | [Gemini 2.5 Flash (April 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Google | 12.58% |
| 34 | [Qwen3 235B-A22B](/models/qwen3-235b-a22b) | Alibaba | 11.75% |
| 35 | [claude-opus-4-1-20250805-thinking](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Anthropic | 11.26% |
| 36 | [Claude Opus 4 (Thinking)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Anthropic | 10.80% |
| 37 | [Gemini 2.5 Flash Preview (May 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Google | 10.72% |
| 38 | [o3 mini (medium)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 10.31% |
| 39 | [GPT-OSS 20B](/models/gpt-oss-20b) | OpenAI | 9.73% |
| 40 | [glm-4p5](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Z.AI | 9.64% |
| 41 | [glm-4p5-air](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Z.AI | 9.41% |
| 42 | [DeepSeek-R1](/models/deepseek-r1) | DeepSeek | 8.54% |
| 43 | [gemini-3.1-flash-lite-preview](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Google | 8.02% |
| 44 | [Claude 3.7 Sonnet (Thinking)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Anthropic | 7.89% |
| 45 | [o1 (December 2024)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 7.75% |
| 46 | [o1-pro](/models/o1-pro) | OpenAI | 7.71% |
| 47 | [Claude Sonnet 4.5](/models/claude-sonnet-4-5) | Anthropic | 7.65% |
| 48 | [Claude Sonnet 4 (Thinking)\n](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Anthropic | 7.60% |
| 49 | [claude-opus-4-1-20250805](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Anthropic | 7.37% |
| 50 | [Gemini 2.0 Flash Thinking (January 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Google | 6.55% |
| 51 | [gpt-5.1-instant](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 6.49% |
| 52 | [Claude Opus 4 \n](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Anthropic | 6.26% |
| 53 | [GPT 4.5 Preview](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 5.80% |
| 54 | [Claude Sonnet 4](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Anthropic | 5.42% |
| 55 | [Llama 4 Maverick](/models/llama-4-maverick) | Meta | 5.34% |
| 56 | [GPT-4.1](/models/gpt-4-1) | OpenAI | 4.97% |
| 57 | [kimi-k2-instruct](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Moonshot AI | 4.68% |
| 58 | [DeepSeek V3 (March 2025)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Deepseek | 4.55% |
| 59 | [Gemini-1.5-Pro-002](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Google | 4.55% |
| 60 | [Nova Micro](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Amazon | 4.41% |
| 61 | [Mistral Medium 3](/models/mistral-medium-3) | Mistral | 4.36% |
| 62 | [Claude 3.5 Sonnet (October 2024)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Anthropic | 4.32% |
| 63 | [Nova Pro](/models/nova-pro) | Amazon | 4.32% |
| 64 | [Nova Lite](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | Amazon | 3.76% |
| 65 | [GPT-4o (November 2024)](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) | OpenAI | 2.32% |

## FAQ

### What does Humanity's Last Exam (Text Only) 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 (Text Only) snapshot?

GPT-6 Astra currently leads the published Humanity's Last Exam (Text Only) snapshot with a score of 54.17%.

### How many models are evaluated on Humanity's Last Exam (Text Only)?

The October 15, 2026 snapshot contains 65 AI models.

### Does Humanity's Last Exam (Text Only) affect BenchLM's overall score?

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