# Chart Reasoning on CharXiv Rea

Source: [Chart Reasoning on CharXiv Rea](<https://www.sota2.com/research/sota/chart-reasoning-on-charxiv-rea>)

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

Data updated: 2026-03-31T08:40:59.271Z

Evaluation of model performance on the CharXiv Rea dataset for Chart Reasoning tasks, measured using 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 | Chart Reasoning |
| Dataset (including any stated split) | CharXiv Rea |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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/chart-reasoning-on-charxiv-rea.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Qwen2.5-VL-7B-Instruct + GRPO + KAWHI | 46.4 | Not recorded | 69cb54fe0636e5bf4afc9e86 |
| Qwen2.5-VL-7B-Instruct + GRPO | 44.3 | Not recorded | 69cb54fe0636e5bf4afc9e86 |
| Qwen2.5-VL-7B-Instruct | 42.1 | Not recorded | 69cb54fe0636e5bf4afc9e86 |

### Qwen2.5-VL-7B-Instruct + GRPO + KAWHI

Accuracy: 46.4. Result ID: 69cb888a47f34ce924b3df9c.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54fe0636e5bf4afc9e86.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb54fe0636e5bf4afc9e86.md>)

Recorded evaluation configuration: {"Training Method":"GRPO + KAWHI"}. This record does not assert independent reproduction.

### Qwen2.5-VL-7B-Instruct + GRPO

Accuracy: 44.3. Result ID: 69cb888a47f34ce924b3df9b.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54fe0636e5bf4afc9e86.md>)

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

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

Recorded evaluation configuration: {"Training Method":"GRPO"}. This record does not assert independent reproduction.

### Qwen2.5-VL-7B-Instruct

Accuracy: 42.1. Result ID: 69cb888a47f34ce924b3df9a.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54fe0636e5bf4afc9e86.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa773a1e209131cea71d3c.md>)

Recorded evaluation configuration: {"Model Size":"7B"}. This record does not assert independent reproduction.

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