# Mathematical Reasoning on AIME 2025 (Pass@1, FS, N/T, D/T, Repro)

Source: [Mathematical Reasoning on AIME 2025 (Pass@1, FS, N/T, D/T, Repro)](<https://www.sota2.com/research/sota/mathematical-reasoning-on-aime-2025-pass-at-1-fs-n-t-d-t-repro>)

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

Data updated: 2026-04-06T08:41:43.446Z

Evaluation of model performance on the American Invitational Mathematics Examination (AIME) 2025 dataset, reporting results across Pass@1 accuracy, Few-Shot (FS), Normalized/Total (N/T), Direct/Total (D/T), and Reproducibility (Repro) metrics.

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 | Mathematical Reasoning |
| Dataset (including any stated split) | AIME 2025 |
| Metric | Pass@1 Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 15 of 15 |
| 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 Accuracy](<https://www.sota2.com/research/sota/mathematical-reasoning-on-aime-2025-pass-at-1-fs-n-t-d-t-repro.md?metric=Pass%401%20Accuracy>): Higher is better; unit: PERCENT.

- [FS Score](<https://www.sota2.com/research/sota/mathematical-reasoning-on-aime-2025-pass-at-1-fs-n-t-d-t-repro.md?metric=FS%20Score>): Higher is better; unit: PERCENT.

- [N/T Ratio](<https://www.sota2.com/research/sota/mathematical-reasoning-on-aime-2025-pass-at-1-fs-n-t-d-t-repro.md?metric=N%2FT%20Ratio>): Higher is better; unit: PERCENT.

- [D/T Ratio](<https://www.sota2.com/research/sota/mathematical-reasoning-on-aime-2025-pass-at-1-fs-n-t-d-t-repro.md?metric=D%2FT%20Ratio>): Higher is better; unit: PERCENT.

- [Reproducibility](<https://www.sota2.com/research/sota/mathematical-reasoning-on-aime-2025-pass-at-1-fs-n-t-d-t-repro.md?metric=Reproducibility>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Pass@1 Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| RED | 72.2 | Not recorded | 69d33dea7d0172140ec735b1 |
| S-GRPO | 71.1 | Not recorded | 69d33dea7d0172140ec735b1 |
| Vanilla | 68.9 | Not recorded | 69d33dea7d0172140ec735b1 |
| DEER | 66.7 | Not recorded | 69d33dea7d0172140ec735b1 |
| RL + LP | 66.7 | Not recorded | 69d33dea7d0172140ec735b1 |
| RED | 63.3 | Not recorded | 69d33dea7d0172140ec735b1 |
| RL + LP | 62.2 | Not recorded | 69d33dea7d0172140ec735b1 |
| S-GRPO | 60 | Not recorded | 69d33dea7d0172140ec735b1 |
| Vanilla | 58.9 | Not recorded | 69d33dea7d0172140ec735b1 |
| DEER | 57.8 | Not recorded | 69d33dea7d0172140ec735b1 |
| RED | 50 | Not recorded | 69d33dea7d0172140ec735b1 |
| DEER | 48.9 | Not recorded | 69d33dea7d0172140ec735b1 |
| S-GRPO | 48.9 | Not recorded | 69d33dea7d0172140ec735b1 |
| Vanilla | 47.8 | Not recorded | 69d33dea7d0172140ec735b1 |
| RL + LP | 45.6 | Not recorded | 69d33dea7d0172140ec735b1 |

### RED

Pass@1 Accuracy: 72.2. Result ID: 69d371c547f34ce924bb5939.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-32B-thinking"}. This record does not assert independent reproduction.

### S-GRPO

Pass@1 Accuracy: 71.1. Result ID: 69d371c547f34ce924bb5938.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78961e209131cea79249.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-32B-thinking"}. This record does not assert independent reproduction.

### Vanilla

Pass@1 Accuracy: 68.9. Result ID: 69d371c547f34ce924bb5935.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-32B-thinking"}. This record does not assert independent reproduction.

### DEER

Pass@1 Accuracy: 66.7. Result ID: 69d371c547f34ce924bb5936.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ba11e209131cea89778.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-32B-thinking"}. This record does not assert independent reproduction.

### RL + LP

Pass@1 Accuracy: 66.7. Result ID: 69d371c547f34ce924bb5937.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78a81e209131cea79a9b.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-32B-thinking"}. This record does not assert independent reproduction.

### RED

Pass@1 Accuracy: 63.3. Result ID: 69d371c547f34ce924bb593e.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

Recorded evaluation configuration: {"Backbone":"R1-Qwen-32B"}. This record does not assert independent reproduction.

### RL + LP

Pass@1 Accuracy: 62.2. Result ID: 69d371c547f34ce924bb593c.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78a81e209131cea79a9b.md>)

Recorded evaluation configuration: {"Backbone":"R1-Qwen-32B"}. This record does not assert independent reproduction.

### S-GRPO

Pass@1 Accuracy: 60. Result ID: 69d371c547f34ce924bb593d.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78961e209131cea79249.md>)

Recorded evaluation configuration: {"Backbone":"R1-Qwen-32B"}. This record does not assert independent reproduction.

### Vanilla

Pass@1 Accuracy: 58.9. Result ID: 69d371c547f34ce924bb593a.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

Recorded evaluation configuration: {"Backbone":"R1-Qwen-32B"}. This record does not assert independent reproduction.

### DEER

Pass@1 Accuracy: 57.8. Result ID: 69d371c547f34ce924bb593b.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ba11e209131cea89778.md>)

Recorded evaluation configuration: {"Backbone":"R1-Qwen-32B"}. This record does not assert independent reproduction.

### RED

Pass@1 Accuracy: 50. Result ID: 69d371c547f34ce924bb5943.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

Recorded evaluation configuration: {"Backbone":"R1-Llama-70B"}. This record does not assert independent reproduction.

### DEER

Pass@1 Accuracy: 48.9. Result ID: 69d371c547f34ce924bb5940.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ba11e209131cea89778.md>)

Recorded evaluation configuration: {"Backbone":"R1-Llama-70B"}. This record does not assert independent reproduction.

### S-GRPO

Pass@1 Accuracy: 48.9. Result ID: 69d371c547f34ce924bb5942.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78961e209131cea79249.md>)

Recorded evaluation configuration: {"Backbone":"R1-Llama-70B"}. This record does not assert independent reproduction.

### Vanilla

Pass@1 Accuracy: 47.8. Result ID: 69d371c547f34ce924bb593f.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

Recorded evaluation configuration: {"Backbone":"R1-Llama-70B"}. This record does not assert independent reproduction.

### RL + LP

Pass@1 Accuracy: 45.6. Result ID: 69d371c547f34ce924bb5941.

[Reporting paper](<https://www.sota2.com/research/paper/69d33dea7d0172140ec735b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78a81e209131cea79a9b.md>)

Recorded evaluation configuration: {"Backbone":"R1-Llama-70B"}. This record does not assert independent reproduction.

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