# Mathematical Reasoning on Mathematical Reasoning Aggregate (Pass@1)

Source: [Mathematical Reasoning on Mathematical Reasoning Aggregate (Pass@1)](<https://www.sota2.com/research/sota/mathematical-reasoning-on-mathematical-reasoning-aggregate-pass-at-1>)

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

Data updated: 2026-06-23T11:35:20.495Z

Evaluation of model performance on the Mathematical Reasoning Aggregate dataset, specifically reporting Pass@1 accuracy and average output token length.

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) | Mathematical Reasoning Aggregate |
| Metric | Average Accuracy (Pass@1) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

- [Average Accuracy (Pass@1)](<https://www.sota2.com/research/sota/mathematical-reasoning-on-mathematical-reasoning-aggregate-pass-at-1.md?metric=Average%20Accuracy%20(Pass%401)>): Higher is better; unit: PERCENT.

- [Average Output Length (Tokens)](<https://www.sota2.com/research/sota/mathematical-reasoning-on-mathematical-reasoning-aggregate-pass-at-1.md?metric=Average%20Output%20Length%20(Tokens)>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Average Accuracy (Pass@1) | Result date | Report paper |
| --- | --- | --- | --- |
| Qwen3-8B+E2C-(SFT+RL) | 61.5 | Not recorded | 68fa7bc71e209131cea8a6a6 |
| Qwen3-8B+GRPO (Baseline) | 60 | Not recorded | 68fa7bc71e209131cea8a6a6 |
| Qwen3-4B+E2C-(SFT+RL) | 58.4 | Not recorded | 68fa7bc71e209131cea8a6a6 |
| Qwen3-8B+E2C-SFT+GRPO | 57.7 | Not recorded | 68fa7bc71e209131cea8a6a6 |
| Qwen3-4B+GRPO (Baseline) | 56.5 | Not recorded | 68fa7bc71e209131cea8a6a6 |
| Qwen3-4B+E2C-SFT+GRPO | 55.2 | Not recorded | 68fa7bc71e209131cea8a6a6 |
| Qwen3-8B+E2C-SFT | 46.8 | Not recorded | 68fa7bc71e209131cea8a6a6 |
| Qwen3-8B+Prompt (Zero-shot) | 46.6 | Not recorded | 68fa7bc71e209131cea8a6a6 |

### Qwen3-8B+E2C-(SFT+RL)

Average Accuracy (Pass@1): 61.5. Result ID: 6a3a6f60fbc373592c37f1e9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

Recorded evaluation configuration: {"Model Series":"Qwen3 8B","Training Stage":"E2C-(SFT+RL)"}. This record does not assert independent reproduction.

### Qwen3-8B+GRPO (Baseline)

Average Accuracy (Pass@1): 60. Result ID: 6a3a6f60fbc373592c37f1e7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75001e209131cea652ab.md>)

Recorded evaluation configuration: {"Model Series":"Qwen3 8B","Training Stage":"GRPO"}. This record does not assert independent reproduction.

### Qwen3-4B+E2C-(SFT+RL)

Average Accuracy (Pass@1): 58.4. Result ID: 6a3a6f60fbc373592c37f1ec.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

Recorded evaluation configuration: {"Model Series":"Qwen3 4B","Training Stage":"E2C-(SFT+RL)"}. This record does not assert independent reproduction.

### Qwen3-8B+E2C-SFT+GRPO

Average Accuracy (Pass@1): 57.7. Result ID: 6a3a6f60fbc373592c37f1e8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

Recorded evaluation configuration: {"Model Series":"Qwen3 8B","Training Stage":"E2C-SFT+GRPO"}. This record does not assert independent reproduction.

### Qwen3-4B+GRPO (Baseline)

Average Accuracy (Pass@1): 56.5. Result ID: 6a3a6f60fbc373592c37f1ea.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75001e209131cea652ab.md>)

Recorded evaluation configuration: {"Model Series":"Qwen3 4B","Training Stage":"GRPO"}. This record does not assert independent reproduction.

### Qwen3-4B+E2C-SFT+GRPO

Average Accuracy (Pass@1): 55.2. Result ID: 6a3a6f60fbc373592c37f1eb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

Recorded evaluation configuration: {"Model Series":"Qwen3 4B","Training Stage":"E2C-SFT+GRPO"}. This record does not assert independent reproduction.

### Qwen3-8B+E2C-SFT

Average Accuracy (Pass@1): 46.8. Result ID: 6a3a6f60fbc373592c37f1ee.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

Recorded evaluation configuration: {"Model Series":"Qwen3 8B","Training Stage":"E2C-SFT"}. This record does not assert independent reproduction.

### Qwen3-8B+Prompt (Zero-shot)

Average Accuracy (Pass@1): 46.6. Result ID: 6a3a6f60fbc373592c37f1ed.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6a6.md>)

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

Recorded evaluation configuration: {"Model Series":"Qwen3 8B","Evaluation Mode":"Zero-shot"}. This record does not assert independent reproduction.

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