# Geometric Reasoning on Geometry3K (avg@8 accuracy)

Source: [Geometric Reasoning on Geometry3K (avg@8 accuracy)](<https://www.sota2.com/research/sota/geometric-reasoning-on-geometry3k-avg-at-8-accuracy>)

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

Data updated: 2026-04-15T09:05:14.806Z

Evaluation of model performance on geometric reasoning tasks using the Geometry3K dataset, measured by 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 | Geometric Reasoning |
| Dataset (including any stated split) | Geometry3K |
| 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/geometric-reasoning-on-geometry3k-avg-at-8-accuracy.md?metric=Avg%408%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Avg@8 Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| PAPO\_D | 44.11 | Not recorded | 68fa7a761e209131cea8363f |
| PAPO\_G | 41.08 | Not recorded | 68fa7a761e209131cea8363f |
| PAPO\_G | 40.25 | Not recorded | 68fa7a761e209131cea8363f |
| GRPO | 40.18 | Not recorded | 68fa7a761e209131cea8363f |
| GRPO | 39.29 | Not recorded | 68fa7a761e209131cea8363f |
| DAPO | 35.92 | Not recorded | 68fa7a761e209131cea8363f |
| PAPO\_D | 35.65 | Not recorded | 68fa7a761e209131cea8363f |
| DAPO | 31.2 | Not recorded | 68fa7a761e209131cea8363f |
| PAPO\_G | 30.95 | Not recorded | 68fa7a761e209131cea8363f |
| GRPO | 28.72 | Not recorded | 68fa7a761e209131cea8363f |

### PAPO\_D

Avg@8 Accuracy: 44.11. Result ID: 69df54ba47f34ce924c84426.

[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","Training Algorithm":"PAPO\_D","Inference Temperature":"1.0"}. This record does not assert independent reproduction.

### PAPO\_G

Avg@8 Accuracy: 41.08. Result ID: 69df54ba47f34ce924c84428.

[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","Training Algorithm":"PAPO\_G","Inference Temperature":"1.0"}. This record does not assert independent reproduction.

### PAPO\_G

Avg@8 Accuracy: 40.25. Result ID: 69df54ba47f34ce924c84422.

[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","Training Algorithm":"PAPO\_G","Inference Temperature":"1.0"}. This record does not assert independent reproduction.

### GRPO

Avg@8 Accuracy: 40.18. Result ID: 69df54ba47f34ce924c84421.

[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","Training Algorithm":"GRPO","Inference Temperature":"1.0"}. This record does not assert independent reproduction.

### GRPO

Avg@8 Accuracy: 39.29. Result ID: 69df54ba47f34ce924c84427.

[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","Training Algorithm":"GRPO","Inference Temperature":"1.0"}. This record does not assert independent reproduction.

### DAPO

Avg@8 Accuracy: 35.92. Result ID: 69df54ba47f34ce924c84425.

[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","Training Algorithm":"DAPO","Inference Temperature":"1.0"}. This record does not assert independent reproduction.

### PAPO\_D

Avg@8 Accuracy: 35.65. Result ID: 69df54ba47f34ce924c84424.

[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","Training Algorithm":"PAPO\_D","Inference Temperature":"1.0"}. This record does not assert independent reproduction.

### DAPO

Avg@8 Accuracy: 31.2. Result ID: 69df54ba47f34ce924c84423.

[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","Training Algorithm":"DAPO","Inference Temperature":"1.0"}. This record does not assert independent reproduction.

### PAPO\_G

Avg@8 Accuracy: 30.95. Result ID: 69df54ba47f34ce924c84420.

[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","Training Algorithm":"PAPO\_G","Inference Temperature":"1.0"}. This record does not assert independent reproduction.

### GRPO

Avg@8 Accuracy: 28.72. Result ID: 69df54ba47f34ce924c8441f.

[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","Training Algorithm":"GRPO","Inference Temperature":"1.0"}. This record does not assert independent reproduction.

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