# Question Answering on GPQA Diamond (Accuracy and Tokens)

Source: [Question Answering on GPQA Diamond (Accuracy and Tokens)](<https://www.sota2.com/research/sota/question-answering-on-gpqa-diamond-accuracy-and-tokens>)

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

Data updated: 2026-04-27T07:40:17.425Z

Evaluation of model performance on the GPQA Diamond dataset for Question Answering, measured by Accuracy and Token usage.

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 | Question Answering |
| Dataset (including any stated split) | GPQA Diamond |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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/question-answering-on-gpqa-diamond-accuracy-and-tokens.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [Tokens](<https://www.sota2.com/research/sota/question-answering-on-gpqa-diamond-accuracy-and-tokens.md?metric=Tokens>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| SFT + RL | 51.5 | Not recorded | 69eeeda455a3c32af89f9868 |
| Abstract-CoT (Warm-up + RL) | 50.5 | Not recorded | 69eeeda455a3c32af89f9868 |
| SFT (CoT) | 45.5 | Not recorded | 69eeeda455a3c32af89f9868 |
| Baseline | 44.9 | Not recorded | 69eeeda455a3c32af89f9868 |
| Abstract-CoT (Warm-up) | 44.4 | Not recorded | 69eeeda455a3c32af89f9868 |
| Abstract-CoT (RL-only) | 41.9 | Not recorded | 69eeeda455a3c32af89f9868 |
| SFT (no CoT) | 39.4 | Not recorded | 69eeeda455a3c32af89f9868 |

### SFT + RL

Accuracy: 51.5. Result ID: 69ef12c547f34ce924d4f691.

[Reporting paper](<https://www.sota2.com/research/paper/69eeeda455a3c32af89f9868.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73101e209131cea5a395.md>)

[Implementation](<https://github.com/deepseek-ai/DeepSeek-Math>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### Abstract-CoT (Warm-up + RL)

Accuracy: 50.5. Result ID: 69ef12c547f34ce924d4f694.

[Reporting paper](<https://www.sota2.com/research/paper/69eeeda455a3c32af89f9868.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eeeda455a3c32af89f9868.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### SFT (CoT)

Accuracy: 45.5. Result ID: 69ef12c547f34ce924d4f690.

[Reporting paper](<https://www.sota2.com/research/paper/69eeeda455a3c32af89f9868.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### Baseline

Accuracy: 44.9. Result ID: 69ef12c547f34ce924d4f68e.

[Reporting paper](<https://www.sota2.com/research/paper/69eeeda455a3c32af89f9868.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### Abstract-CoT (Warm-up)

Accuracy: 44.4. Result ID: 69ef12c547f34ce924d4f693.

[Reporting paper](<https://www.sota2.com/research/paper/69eeeda455a3c32af89f9868.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eeeda455a3c32af89f9868.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### Abstract-CoT (RL-only)

Accuracy: 41.9. Result ID: 69ef12c547f34ce924d4f692.

[Reporting paper](<https://www.sota2.com/research/paper/69eeeda455a3c32af89f9868.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eeeda455a3c32af89f9868.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### SFT (no CoT)

Accuracy: 39.4. Result ID: 69ef12c547f34ce924d4f68f.

[Reporting paper](<https://www.sota2.com/research/paper/69eeeda455a3c32af89f9868.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

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