# Vision-Dependent Multimodal Math Reasoning on MathVerseV Vision-centric

Source: [Vision-Dependent Multimodal Math Reasoning on MathVerseV Vision-centric](<https://www.sota2.com/research/sota/vision-dependent-multimodal-math-reasoning-on-mathversev-vision-centric>)

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

Data updated: 2026-04-15T09:12:37.802Z

Evaluation of model performance on the MathVerseV Vision-centric dataset for Vision-Dependent Multimodal Math Reasoning, measured using avg@8 accuracy.

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 | Vision-Dependent Multimodal Math Reasoning |
| Dataset (including any stated split) | MathVerseV Vision-centric |
| Metric | Avg@8 Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 10 of 10 |
| 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

- [Avg@8 Accuracy](<https://www.sota2.com/research/sota/vision-dependent-multimodal-math-reasoning-on-mathversev-vision-centric.md?metric=Avg%408%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Avg@8 Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| PAPO\_G | 64.97 | Not recorded | 68fa7a761e209131cea8363f |
| PAPO\_D | 64.87 | Not recorded | 68fa7a761e209131cea8363f |
| GRPO | 61.71 | Not recorded | 68fa7a761e209131cea8363f |
| PAPO\_D | 57.72 | Not recorded | 68fa7a761e209131cea8363f |
| PAPO\_G | 53.95 | Not recorded | 68fa7a761e209131cea8363f |
| DAPO | 53.09 | Not recorded | 68fa7a761e209131cea8363f |
| GRPO | 52.26 | Not recorded | 68fa7a761e209131cea8363f |
| DAPO | 51.04 | Not recorded | 68fa7a761e209131cea8363f |
| PAPO\_G | 50.05 | Not recorded | 68fa7a761e209131cea8363f |
| GRPO | 45.41 | Not recorded | 68fa7a761e209131cea8363f |

### PAPO\_G

Avg@8 Accuracy: 64.97. Result ID: 69df54ba47f34ce924c8447c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL","Model Size":"7B"}. This record does not assert independent reproduction.

### PAPO\_D

Avg@8 Accuracy: 64.87. Result ID: 69df54ba47f34ce924c84480.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL","Model Size":"7B"}. This record does not assert independent reproduction.

### GRPO

Avg@8 Accuracy: 61.71. Result ID: 69df54ba47f34ce924c8447b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

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

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

Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL","Model Size":"7B"}. This record does not assert independent reproduction.

### PAPO\_D

Avg@8 Accuracy: 57.72. Result ID: 69df54ba47f34ce924c8447e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL","Model Size":"3B"}. This record does not assert independent reproduction.

### PAPO\_G

Avg@8 Accuracy: 53.95. Result ID: 69df54ba47f34ce924c8447a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL","Model Size":"3B"}. This record does not assert independent reproduction.

### DAPO

Avg@8 Accuracy: 53.09. Result ID: 69df54ba47f34ce924c8447d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78a81e209131cea79ac6.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL","Model Size":"3B"}. This record does not assert independent reproduction.

### GRPO

Avg@8 Accuracy: 52.26. Result ID: 69df54ba47f34ce924c84479.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

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

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

Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL","Model Size":"3B"}. This record does not assert independent reproduction.

### DAPO

Avg@8 Accuracy: 51.04. Result ID: 69df54ba47f34ce924c8447f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78a81e209131cea79ac6.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL","Model Size":"7B"}. This record does not assert independent reproduction.

### PAPO\_G

Avg@8 Accuracy: 50.05. Result ID: 69df54ba47f34ce924c84482.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-VL (thinking)","Model Size":"2B"}. This record does not assert independent reproduction.

### GRPO

Avg@8 Accuracy: 45.41. Result ID: 69df54ba47f34ce924c84481.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a761e209131cea8363f.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-VL (thinking)","Model Size":"2B"}. This record does not assert independent reproduction.

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