# Mathematical Reasoning on GSM8K (full set)

Source: [Mathematical Reasoning on GSM8K (full set)](<https://www.sota2.com/research/sota/mathematical-reasoning-on-gsm8k-full-set>)

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

Data updated: 2026-05-11T10:03:04.671Z

Evaluation of model performance on the full set (1,319 samples) of the GSM8K dataset, focusing on mathematical reasoning capabilities, typically measured by 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) | GSM8K n=1,319 (full set) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| 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/mathematical-reasoning-on-gsm8k-full-set.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [Avg Output Tokens](<https://www.sota2.com/research/sota/mathematical-reasoning-on-gsm8k-full-set.md?metric=Avg%20Output%20Tokens>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| IRIS-single | 91.43 | Not recorded | 6a0162b8dbc836b87d8232fd |
| Mrsd | 90.9 | Not recorded | 6a0162b8dbc836b87d8232fd |
| Nothink | 87.49 | Not recorded | 6a0162b8dbc836b87d8232fd |
| Town | 85.97 | Not recorded | 6a0162b8dbc836b87d8232fd |

### IRIS-single

Accuracy: 91.43. Result ID: 6a01a95847f34ce924e40f73.

[Reporting paper](<https://www.sota2.com/research/paper/6a0162b8dbc836b87d8232fd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0162b8dbc836b87d8232fd.md>)

Recorded evaluation configuration: {"Budget":"256/512/128","Model":"Qwen3-8B","Seed":"42"}. This record does not assert independent reproduction.

### Mrsd

Accuracy: 90.9. Result ID: 6a01a95847f34ce924e40f74.

[Reporting paper](<https://www.sota2.com/research/paper/6a0162b8dbc836b87d8232fd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0162b8dbc836b87d8232fd.md>)

Recorded evaluation configuration: {"Rounds":"3-round","Model":"Qwen3-8B","Seed":"42"}. This record does not assert independent reproduction.

### Nothink

Accuracy: 87.49. Result ID: 6a01a95847f34ce924e40f71.

[Reporting paper](<https://www.sota2.com/research/paper/6a0162b8dbc836b87d8232fd.md>)

Recorded evaluation configuration: {"Budget":"256","Model":"Qwen3-8B","Seed":"42"}. This record does not assert independent reproduction.

### Town

Accuracy: 85.97. Result ID: 6a01a95847f34ce924e40f72.

[Reporting paper](<https://www.sota2.com/research/paper/6a0162b8dbc836b87d8232fd.md>)

Recorded evaluation configuration: {"Budget":"256→512","Model":"Qwen3-8B","Seed":"42"}. This record does not assert independent reproduction.

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