# Visual Mathematical Reasoning on MathVista (BoN@8)

Source: [Visual Mathematical Reasoning on MathVista (BoN@8)](<https://www.sota2.com/research/sota/visual-mathematical-reasoning-on-mathvista-bon-at-8>)

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

Data updated: 2026-03-25T13:41:00.361Z

Evaluation of model performance on the MathVista dataset for Visual Mathematical Reasoning, specifically reporting metrics derived from the BoN@8 sampling strategy.

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 | Visual Mathematical Reasoning |
| Dataset (including any stated split) | MathVista |
| Metric | BoN@8 Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 15 of 15 |
| 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

- [BoN@8 Accuracy](<https://www.sota2.com/research/sota/visual-mathematical-reasoning-on-mathvista-bon-at-8.md?metric=BoN%408%20Accuracy>): Higher is better; unit: PERCENT.

- [Delta (BoN@8 - Pass@1)](<https://www.sota2.com/research/sota/visual-mathematical-reasoning-on-mathvista-bon-at-8.md?metric=Delta%20(BoN%408%20-%20Pass%401)>): Higher is better; unit: SCALAR.

## Recorded results

| Method | BoN@8 Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| InternVL2.5-38B + EVPV-PRM | 83.5 | Not recorded | 69ba31af0636e5bf4afc0a3c |
| InternVL2.5-26B + EVPV-PRM | 79.6 | Not recorded | 69ba31af0636e5bf4afc0a3c |
| InternVL2.5-8B + EVPV-PRM | 76.3 | Not recorded | 69ba31af0636e5bf4afc0a3c |
| InternVL2.5-38B + VisualPRM | 73.9 | Not recorded | 69ba31af0636e5bf4afc0a3c |
| InternVL2.5-26B + VisualPRM | 73.1 | Not recorded | 69ba31af0636e5bf4afc0a3c |
| InternVL2.5-38B | 71.9 | Not recorded | 69ba31af0636e5bf4afc0a3c |
| Gemini-2.0-Flash | 70.4 | Not recorded | 69ba31af0636e5bf4afc0a3c |
| InternVL2.5-8B + VisualPRM | 68.5 | Not recorded | 69ba31af0636e5bf4afc0a3c |
| InternVL2.5-26B | 68.2 | Not recorded | 69ba31af0636e5bf4afc0a3c |
| Claude-3.5-Sonnet | 65.3 | Not recorded | 69ba31af0636e5bf4afc0a3c |
| InternVL2.5-8B | 64.5 | Not recorded | 69ba31af0636e5bf4afc0a3c |
| PEPO\_G | 63.48 | Not recorded | 69c36c3b0636e5bf4afc662b |
| PAPO\_G | 61.38 | Not recorded | 69c36c3b0636e5bf4afc662b |
| GPT-4o | 60 | Not recorded | 69ba31af0636e5bf4afc0a3c |
| GRPO | 59.34 | Not recorded | 69c36c3b0636e5bf4afc662b |

### InternVL2.5-38B + EVPV-PRM

BoN@8 Accuracy: 83.5. Result ID: 69ba613447f34ce924a438da.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

Recorded evaluation configuration: {"Policy Model":"InternVL2.5-38B","Process Reward Model (PRM)":"EVPV-PRM","Reranking Strategy":"Best-of-8"}. This record does not assert independent reproduction.

### InternVL2.5-26B + EVPV-PRM

BoN@8 Accuracy: 79.6. Result ID: 69ba613447f34ce924a438d7.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

Recorded evaluation configuration: {"Policy Model":"InternVL2.5-26B","Process Reward Model (PRM)":"EVPV-PRM","Reranking Strategy":"Best-of-8"}. This record does not assert independent reproduction.

### InternVL2.5-8B + EVPV-PRM

BoN@8 Accuracy: 76.3. Result ID: 69ba613447f34ce924a438d4.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

Recorded evaluation configuration: {"Policy Model":"InternVL2.5-8B","Process Reward Model (PRM)":"EVPV-PRM","Reranking Strategy":"Best-of-8"}. This record does not assert independent reproduction.

### InternVL2.5-38B + VisualPRM

BoN@8 Accuracy: 73.9. Result ID: 69ba613447f34ce924a438d9.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

Recorded evaluation configuration: {"Policy Model":"InternVL2.5-38B","Process Reward Model (PRM)":"VisualPRM","Reranking Strategy":"Best-of-8"}. This record does not assert independent reproduction.

### InternVL2.5-26B + VisualPRM

BoN@8 Accuracy: 73.1. Result ID: 69ba613447f34ce924a438d6.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

Recorded evaluation configuration: {"Policy Model":"InternVL2.5-26B","Process Reward Model (PRM)":"VisualPRM","Reranking Strategy":"Best-of-8"}. This record does not assert independent reproduction.

### InternVL2.5-38B

BoN@8 Accuracy: 71.9. Result ID: 69ba613447f34ce924a438d8.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

Recorded evaluation configuration: {"Policy Model":"InternVL2.5-38B","Process Reward Model (PRM)":"None","Reranking Strategy":"Best-of-8"}. This record does not assert independent reproduction.

### Gemini-2.0-Flash

BoN@8 Accuracy: 70.4. Result ID: 69ba613447f34ce924a438d0.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

Recorded evaluation configuration: {"Reranking Strategy":"Best-of-8"}. This record does not assert independent reproduction.

### InternVL2.5-8B + VisualPRM

BoN@8 Accuracy: 68.5. Result ID: 69ba613447f34ce924a438d3.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

Recorded evaluation configuration: {"Policy Model":"InternVL2.5-8B","Process Reward Model (PRM)":"VisualPRM","Reranking Strategy":"Best-of-8"}. This record does not assert independent reproduction.

### InternVL2.5-26B

BoN@8 Accuracy: 68.2. Result ID: 69ba613447f34ce924a438d5.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

Recorded evaluation configuration: {"Policy Model":"InternVL2.5-26B","Process Reward Model (PRM)":"None","Reranking Strategy":"Best-of-8"}. This record does not assert independent reproduction.

### Claude-3.5-Sonnet

BoN@8 Accuracy: 65.3. Result ID: 69ba613447f34ce924a438d1.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

Recorded evaluation configuration: {"Reranking Strategy":"Best-of-8"}. This record does not assert independent reproduction.

### InternVL2.5-8B

BoN@8 Accuracy: 64.5. Result ID: 69ba613447f34ce924a438d2.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

Recorded evaluation configuration: {"Policy Model":"InternVL2.5-8B","Process Reward Model (PRM)":"None","Reranking Strategy":"Best-of-8"}. This record does not assert independent reproduction.

### PEPO\_G

BoN@8 Accuracy: 63.48. Result ID: 69c3e5e847f34ce924ad7998.

[Reporting paper](<https://www.sota2.com/research/paper/69c36c3b0636e5bf4afc662b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69c36c3b0636e5bf4afc662b.md>)

[Implementation](<https://github.com/xzxxntxdy/PEPO>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-3B-Instruct","Training Data":"ViRL39K"}. This record does not assert independent reproduction.

### PAPO\_G

BoN@8 Accuracy: 61.38. Result ID: 69c3e5e847f34ce924ad7997.

[Reporting paper](<https://www.sota2.com/research/paper/69c36c3b0636e5bf4afc662b.md>)

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

Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-3B-Instruct","Training Data":"ViRL39K"}. This record does not assert independent reproduction.

### GPT-4o

BoN@8 Accuracy: 60. Result ID: 69ba613447f34ce924a438cf.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31af0636e5bf4afc0a3c.md>)

Recorded evaluation configuration: {"Reranking Strategy":"Best-of-8"}. This record does not assert independent reproduction.

### GRPO

BoN@8 Accuracy: 59.34. Result ID: 69c3e5e847f34ce924ad7996.

[Reporting paper](<https://www.sota2.com/research/paper/69c36c3b0636e5bf4afc662b.md>)

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

Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-3B-Instruct","Training Data":"ViRL39K"}. This record does not assert independent reproduction.

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