# Mathematical Reasoning on ViRL39K (test)

Source: [Mathematical Reasoning on ViRL39K (test)](<https://www.sota2.com/research/sota/mathematical-reasoning-on-virl39k-test>)

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

Data updated: 2026-04-27T06:49:33.289Z

Evaluation of model performance on the test split of the ViRL39K dataset, focusing on mathematical reasoning capabilities across various sub-topics including coding, geometry, and STEM knowledge.

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) | ViRL39K (test) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 2 of 2 |
| 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-virl39k-test.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [Code Accuracy](<https://www.sota2.com/research/sota/mathematical-reasoning-on-virl39k-test.md?metric=Code%20Accuracy>): Higher is better; unit: PERCENT.

- [Correctness Rate](<https://www.sota2.com/research/sota/mathematical-reasoning-on-virl39k-test.md?metric=Correctness%20Rate>): Higher is better; unit: PERCENT.

- [GradeSchool Geometric Accuracy](<https://www.sota2.com/research/sota/mathematical-reasoning-on-virl39k-test.md?metric=GradeSchool%20Geometric%20Accuracy>): Higher is better; unit: PERCENT.

- [GradeSchool Non-Geo Math Accuracy](<https://www.sota2.com/research/sota/mathematical-reasoning-on-virl39k-test.md?metric=GradeSchool%20Non-Geo%20Math%20Accuracy>): Higher is better; unit: PERCENT.

- [GradeSchool Science Accuracy](<https://www.sota2.com/research/sota/mathematical-reasoning-on-virl39k-test.md?metric=GradeSchool%20Science%20Accuracy>): Higher is better; unit: PERCENT.

- [Broader STEM Topics Accuracy](<https://www.sota2.com/research/sota/mathematical-reasoning-on-virl39k-test.md?metric=Broader%20STEM%20Topics%20Accuracy>): Higher is better; unit: PERCENT.

- [Social Science Accuracy](<https://www.sota2.com/research/sota/mathematical-reasoning-on-virl39k-test.md?metric=Social%20Science%20Accuracy>): Higher is better; unit: PERCENT.

- [Spatial Reasoning Accuracy](<https://www.sota2.com/research/sota/mathematical-reasoning-on-virl39k-test.md?metric=Spatial%20Reasoning%20Accuracy>): Higher is better; unit: PERCENT.

- [Tables/Diagrams/Charts Accuracy](<https://www.sota2.com/research/sota/mathematical-reasoning-on-virl39k-test.md?metric=Tables%2FDiagrams%2FCharts%20Accuracy>): Higher is better; unit: PERCENT.

- [Overall Accuracy](<https://www.sota2.com/research/sota/mathematical-reasoning-on-virl39k-test.md?metric=Overall%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Base model | 18.06 | Not recorded | 69eeed8a55a3c32af89f970d |
| GRPO-based RL | 8.33 | Not recorded | 69eeed8a55a3c32af89f970d |

### Base model

Accuracy: 18.06. Result ID: 69ef061847f34ce924d4972c.

[Reporting paper](<https://www.sota2.com/research/paper/69eeed8a55a3c32af89f970d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"GRPO-based RL":false,"PromptTokLen (Overall)":"377.10 ± 577.74","OutputTokLen (Overall)":"1031.82 ± 1679.20"}. This record does not assert independent reproduction.

### GRPO-based RL

Accuracy: 8.33. Result ID: 69ef061847f34ce924d4972b.

[Reporting paper](<https://www.sota2.com/research/paper/69eeed8a55a3c32af89f970d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eeed8a55a3c32af89f970d.md>)

Recorded evaluation configuration: {"GRPO-based RL":true,"PromptTokLen (Overall)":"856.10 ± 577.74","OutputTokLen (Overall)":"835.92 ± 1435.58"}. This record does not assert independent reproduction.

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