# Expert-level Question Answering on GPQA Main

Source: [Expert-level Question Answering on GPQA Main](<https://www.sota2.com/research/sota/expert-level-question-answering-on-gpqa-main>)

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

Data updated: 2026-06-01T09:55:00.594Z

Evaluation of model performance on the main split of the GPQA dataset for expert-level question answering, reporting the primary score percentage.

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 Question Answering |
| Dataset (including any stated split) | GPQA Main |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/expert-level-question-answering-on-gpqa-main.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Qwen3.5-4B | 40.2 | Not recorded | 6a1d122fdbd7150506f840ea |
| Qwen3-4B | 36.8 | Not recorded | 6a1d122fdbd7150506f840ea |
| Mellum 2 | 35 | Not recorded | 6a1d122fdbd7150506f840ea |
| Qwen2.5-7B | 34.2 | Not recorded | 6a1d122fdbd7150506f840ea |
| OLMo-3-7B | 27.9 | Not recorded | 6a1d122fdbd7150506f840ea |

### Qwen3.5-4B

Accuracy: 40.2. Result ID: 6a1d566ffbc373592c1706f9.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d122fdbd7150506f840ea.md>)

Recorded evaluation configuration: {"Parameters":"4B"}. This record does not assert independent reproduction.

### Qwen3-4B

Accuracy: 36.8. Result ID: 6a1d566ffbc373592c1706f8.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d122fdbd7150506f840ea.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Parameters":"4B"}. This record does not assert independent reproduction.

### Mellum 2

Accuracy: 35. Result ID: 6a1d566ffbc373592c1706f5.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d122fdbd7150506f840ea.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1d122fdbd7150506f840ea.md>)

Recorded evaluation configuration: {"Parameters":"2.5B/12B"}. This record does not assert independent reproduction.

### Qwen2.5-7B

Accuracy: 34.2. Result ID: 6a1d566ffbc373592c1706f7.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d122fdbd7150506f840ea.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa767e1e209131cea6d6a7.md>)

Recorded evaluation configuration: {"Parameters":"7B"}. This record does not assert independent reproduction.

### OLMo-3-7B

Accuracy: 27.9. Result ID: 6a1d566ffbc373592c1706f6.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d122fdbd7150506f840ea.md>)

[Method origin paper](<https://www.sota2.com/research/paper/694238e845614693628503e0.md>)

Recorded evaluation configuration: {"Parameters":"7B"}. This record does not assert independent reproduction.

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