# Expert-Level Human Knowledge Reasoning on Humanity's Last Exam

Source: [Expert-Level Human Knowledge Reasoning on Humanity's Last Exam](<https://www.sota2.com/research/sota/expert-level-human-knowledge-reasoning-on-humanity-s-last-exam>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-05-19T10:32:17.708Z

Evaluation of model performance on the Humanity's Last Exam dataset, specifically targeting Expert-Level Human Knowledge Reasoning capabilities, measured by Pass@1.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Expert-Level Human Knowledge Reasoning |
| Dataset (including any stated split) | Humanity's Last Exam |
| Metric | Pass@1 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 11 of 11 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Pass@1](<https://www.sota2.com/research/sota/expert-level-human-knowledge-reasoning-on-humanity-s-last-exam.md?metric=Pass%401>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Pass@1 | Result date | Report paper |
| --- | --- | --- | --- |
| Tongyi DeepResearch | 38.3 | Not recorded | 69019f853ef7ccd74e5e3b7f |
| Tongyi DeepResearch | 32.9 | Not recorded | 69019f853ef7ccd74e5e3b7f |
| DeepSeek-V3.1 | 29.8 | Not recorded | 69019f853ef7ccd74e5e3b7f |
| Gemini DeepResearch | 26.9 | Not recorded | 69019f853ef7ccd74e5e3b7f |
| Kimi Researcher | 26.9 | Not recorded | 69019f853ef7ccd74e5e3b7f |
| OpenAI DeepResearch | 26.6 | Not recorded | 69019f853ef7ccd74e5e3b7f |
| OpenAI o3 | 24.9 | Not recorded | 69019f853ef7ccd74e5e3b7f |
| GLM 4.5 | 21.2 | Not recorded | 69019f853ef7ccd74e5e3b7f |
| Claude-4-Sonnet | 20.3 | Not recorded | 69019f853ef7ccd74e5e3b7f |
| Kimi K2 | 18.1 | Not recorded | 69019f853ef7ccd74e5e3b7f |
| OpenAI o4-mini | 17.7 | Not recorded | 69019f853ef7ccd74e5e3b7f |

### Tongyi DeepResearch

Pass@1: 38.3. Result ID: 6a0c3bb2fbc373592c02fba2.

[Reporting paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

Recorded evaluation configuration: {"Agent Type":"DeepResearch Agent","Mode":"Heavy Mode"}. This record does not assert independent reproduction.

### Tongyi DeepResearch

Pass@1: 32.9. Result ID: 6a0c3bb2fbc373592c02fba1.

[Reporting paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

Recorded evaluation configuration: {"Agent Type":"DeepResearch Agent","Configuration":"30B-A3B"}. This record does not assert independent reproduction.

### DeepSeek-V3.1

Pass@1: 29.8. Result ID: 6a0c3bb2fbc373592c02fb9a.

[Reporting paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

Recorded evaluation configuration: {"Agent Type":"LLM-based ReAct Agent"}. This record does not assert independent reproduction.

### Gemini DeepResearch

Pass@1: 26.9. Result ID: 6a0c3bb2fbc373592c02fb9f.

[Reporting paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

Recorded evaluation configuration: {"Agent Type":"DeepResearch Agent"}. This record does not assert independent reproduction.

### Kimi Researcher

Pass@1: 26.9. Result ID: 6a0c3bb2fbc373592c02fba0.

[Reporting paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

Recorded evaluation configuration: {"Agent Type":"DeepResearch Agent"}. This record does not assert independent reproduction.

### OpenAI DeepResearch

Pass@1: 26.6. Result ID: 6a0c3bb2fbc373592c02fb9e.

[Reporting paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

Recorded evaluation configuration: {"Agent Type":"DeepResearch Agent"}. This record does not assert independent reproduction.

### OpenAI o3

Pass@1: 24.9. Result ID: 6a0c3bb2fbc373592c02fb9c.

[Reporting paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

Recorded evaluation configuration: {"Agent Type":"LLM-based ReAct Agent"}. This record does not assert independent reproduction.

### GLM 4.5

Pass@1: 21.2. Result ID: 6a0c3bb2fbc373592c02fb98.

[Reporting paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a891e209131cea83c41.md>)

Recorded evaluation configuration: {"Agent Type":"LLM-based ReAct Agent"}. This record does not assert independent reproduction.

### Claude-4-Sonnet

Pass@1: 20.3. Result ID: 6a0c3bb2fbc373592c02fb9b.

[Reporting paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

Recorded evaluation configuration: {"Agent Type":"LLM-based ReAct Agent"}. This record does not assert independent reproduction.

### Kimi K2

Pass@1: 18.1. Result ID: 6a0c3bb2fbc373592c02fb99.

[Reporting paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a4a1e209131cea823a1.md>)

Recorded evaluation configuration: {"Agent Type":"LLM-based ReAct Agent"}. This record does not assert independent reproduction.

### OpenAI o4-mini

Pass@1: 17.7. Result ID: 6a0c3bb2fbc373592c02fb9d.

[Reporting paper](<https://www.sota2.com/research/paper/69019f853ef7ccd74e5e3b7f.md>)

Recorded evaluation configuration: {"Agent Type":"LLM-based ReAct Agent"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.