# Multimodal Math Reasoning on MathVista In-domain

Source: [Multimodal Math Reasoning on MathVista In-domain](<https://www.sota2.com/research/sota/multimodal-math-reasoning-on-mathvista-in-domain>)

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

Data updated: 2026-02-26T05:54:26.230Z

Evaluation of model performance on the MathVista In-domain split for Multimodal Math Reasoning, measured using GVQA, MVQA, and overall ALL scores.

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 | Multimodal Math Reasoning |
| Dataset (including any stated split) | MathVista In-domain |
| Metric | Overall Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 16 of 16 |
| 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

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

- [GVQA Accuracy](<https://www.sota2.com/research/sota/multimodal-math-reasoning-on-mathvista-in-domain.md?metric=GVQA%20Accuracy>): Higher is better; unit: PERCENT.

- [MVQA Accuracy](<https://www.sota2.com/research/sota/multimodal-math-reasoning-on-mathvista-in-domain.md?metric=MVQA%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Overall Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| ADPO | 65.3 | Not recorded | 695c96e653fdd5c36174346e |
| ADPO | 65 | Not recorded | 695c96e653fdd5c36174346e |
| ADPO | 64.8 | Not recorded | 695c96e653fdd5c36174346e |
| MM-Verifier | 64.1 | Not recorded | 695c96e653fdd5c36174346e |
| R1-VL-7B | 63.5 | Not recorded | 695c96e653fdd5c36174346e |
| GRPO | 63.4 | Not recorded | 695c96e653fdd5c36174346e |
| GRPO | 63.4 | Not recorded | 695c96e653fdd5c36174346e |
| GRPO | 62.9 | Not recorded | 695c96e653fdd5c36174346e |
| MM-Verifier | 62.5 | Not recorded | 695c96e653fdd5c36174346e |
| ADPO | 62.4 | Not recorded | 695c96e653fdd5c36174346e |
| GRPO | 62.2 | Not recorded | 695c96e653fdd5c36174346e |
| Base (Qwen2-VL-7B) | 60.7 | Not recorded | 695c96e653fdd5c36174346e |
| Base (Qwen2-VL-7B) | 60.1 | Not recorded | 695c96e653fdd5c36174346e |
| MM-Verifier | 59.8 | Not recorded | 695c96e653fdd5c36174346e |
| Base (Qwen2-VL-7B) | 58.2 | Not recorded | 695c96e653fdd5c36174346e |
| Base (Qwen2-VL-7B) | 57.9 | Not recorded | 695c96e653fdd5c36174346e |

### ADPO

Overall Accuracy: 65.3. Result ID: 6997ecc95698665ec445feb0.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

Recorded evaluation configuration: {"Sample size":12}. This record does not assert independent reproduction.

### ADPO

Overall Accuracy: 65. Result ID: 6997ecc95698665ec445feac.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

Recorded evaluation configuration: {"Sample size":8}. This record does not assert independent reproduction.

### ADPO

Overall Accuracy: 64.8. Result ID: 6997ecc95698665ec445fea8.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

Recorded evaluation configuration: {"Sample size":4}. This record does not assert independent reproduction.

### MM-Verifier

Overall Accuracy: 64.1. Result ID: 6997ecc95698665ec445fead.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa770e1e209131cea70d2d.md>)

[Implementation](<https://github.com/Aurora-slz/MM-Verify>)

Recorded evaluation configuration: {"Sample size":12}. This record does not assert independent reproduction.

### R1-VL-7B

Overall Accuracy: 63.5. Result ID: 6997ecc95698665ec445fea1.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a651e209131cea82e9b.md>)

Recorded evaluation configuration: {"Sample size":1}. This record does not assert independent reproduction.

### GRPO

Overall Accuracy: 63.4. Result ID: 6997ecc95698665ec445fea7.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

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

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

Recorded evaluation configuration: {"Sample size":4}. This record does not assert independent reproduction.

### GRPO

Overall Accuracy: 63.4. Result ID: 6997ecc95698665ec445feaf.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

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

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

Recorded evaluation configuration: {"Sample size":12}. This record does not assert independent reproduction.

### GRPO

Overall Accuracy: 62.9. Result ID: 6997ecc95698665ec445feab.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

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

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

Recorded evaluation configuration: {"Sample size":8}. This record does not assert independent reproduction.

### MM-Verifier

Overall Accuracy: 62.5. Result ID: 6997ecc95698665ec445fea9.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa770e1e209131cea70d2d.md>)

[Implementation](<https://github.com/Aurora-slz/MM-Verify>)

Recorded evaluation configuration: {"Sample size":8}. This record does not assert independent reproduction.

### ADPO

Overall Accuracy: 62.4. Result ID: 6997ecc95698665ec445fea4.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

Recorded evaluation configuration: {"Sample size":1}. This record does not assert independent reproduction.

### GRPO

Overall Accuracy: 62.2. Result ID: 6997ecc95698665ec445fea3.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

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

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

Recorded evaluation configuration: {"Sample size":1}. This record does not assert independent reproduction.

### Base (Qwen2-VL-7B)

Overall Accuracy: 60.7. Result ID: 6997ecc95698665ec445feae.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75151e209131cea65976.md>)

[Implementation](<https://github.com/QwenLM/Qwen2-VL>)

Recorded evaluation configuration: {"Sample size":12}. This record does not assert independent reproduction.

### Base (Qwen2-VL-7B)

Overall Accuracy: 60.1. Result ID: 6997ecc95698665ec445feaa.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75151e209131cea65976.md>)

[Implementation](<https://github.com/QwenLM/Qwen2-VL>)

Recorded evaluation configuration: {"Sample size":8}. This record does not assert independent reproduction.

### MM-Verifier

Overall Accuracy: 59.8. Result ID: 6997ecc95698665ec445fea5.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa770e1e209131cea70d2d.md>)

[Implementation](<https://github.com/Aurora-slz/MM-Verify>)

Recorded evaluation configuration: {"Sample size":4}. This record does not assert independent reproduction.

### Base (Qwen2-VL-7B)

Overall Accuracy: 58.2. Result ID: 6997ecc95698665ec445fea6.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75151e209131cea65976.md>)

[Implementation](<https://github.com/QwenLM/Qwen2-VL>)

Recorded evaluation configuration: {"Sample size":4}. This record does not assert independent reproduction.

### Base (Qwen2-VL-7B)

Overall Accuracy: 57.9. Result ID: 6997ecc95698665ec445fea2.

[Reporting paper](<https://www.sota2.com/research/paper/695c96e653fdd5c36174346e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75151e209131cea65976.md>)

[Implementation](<https://github.com/QwenLM/Qwen2-VL>)

Recorded evaluation configuration: {"Sample size":1}. This record does not assert independent reproduction.

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