# Medical Visual Question Answering on VQA-RAD 10

Source: [Medical Visual Question Answering on VQA-RAD 10](<https://www.sota2.com/research/sota/medical-visual-question-answering-on-vqa-rad-10>)

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

Data updated: 2026-06-19T08:49:06.515Z

Evaluation of model performance on the VQA-RAD 10 dataset for medical visual question answering, reporting F1, Open Accuracy, and Closed Accuracy metrics.

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 | Medical Visual Question Answering |
| Dataset (including any stated split) | VQA-RAD 10 |
| Metric | Closed Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 12 of 12 |
| 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

- [Closed Accuracy](<https://www.sota2.com/research/sota/medical-visual-question-answering-on-vqa-rad-10.md?metric=Closed%20Accuracy>): Higher is better; unit: PERCENT.

- [F1 Score](<https://www.sota2.com/research/sota/medical-visual-question-answering-on-vqa-rad-10.md?metric=F1%20Score>): Higher is better; unit: PERCENT.

- [Open Accuracy (LLM)](<https://www.sota2.com/research/sota/medical-visual-question-answering-on-vqa-rad-10.md?metric=Open%20Accuracy%20(LLM)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Closed Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Med-Gemini | 78.8 | Not recorded | 6a34cd57206542251d7c61fa |
| MedG. (4B) | 69.1 | Not recorded | 6a34cd57206542251d7c61fa |
| RadGrounder | 68.1 | Not recorded | 6a34cd57206542251d7c61fa |
| RadGrounder | 65.9 | Not recorded | 6a34cd57206542251d7c61fa |
| RadGrounder | 64.7 | Not recorded | 6a34cd57206542251d7c61fa |
| RadGrounder | 64.5 | Not recorded | 6a34cd57206542251d7c61fa |
| RadGrounder | 63.2 | Not recorded | 6a34cd57206542251d7c61fa |
| RadGrounder | 62.7 | Not recorded | 6a34cd57206542251d7c61fa |
| PaliG. 2 (3B) | 55.6 | Not recorded | 6a34cd57206542251d7c61fa |
| RadGrounder | 52.9 | Not recorded | 6a34cd57206542251d7c61fa |
| RadGrounder | 52.9 | Not recorded | 6a34cd57206542251d7c61fa |
| Gemma 3 (4B) | 33.6 | Not recorded | 6a34cd57206542251d7c61fa |

### Med-Gemini

Closed Accuracy: 78.8. Result ID: 6a350275fbc373592c32841b.

[Reporting paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73141e209131cea5a5c0.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### MedG. (4B)

Closed Accuracy: 69.1. Result ID: 6a350275fbc373592c32841a.

[Reporting paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79ff1e209131cea808a4.md>)

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

### RadGrounder

Closed Accuracy: 68.1. Result ID: 6a350275fbc373592c328421.

[Reporting paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

Recorded evaluation configuration: {"Training Data":"Report, VQA, Detect."}. This record does not assert independent reproduction.

### RadGrounder

Closed Accuracy: 65.9. Result ID: 6a350275fbc373592c32841f.

[Reporting paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

Recorded evaluation configuration: {"Training Data":"Report, VQA, Slake, VQA-RAD, Detect."}. This record does not assert independent reproduction.

### RadGrounder

Closed Accuracy: 64.7. Result ID: 6a350275fbc373592c32841c.

[Reporting paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

Recorded evaluation configuration: {"Training Data":"Report, VQA, Slake, VQA-RAD, Detect.","Vision Encoder":"Frozen SigLIP (adapted)"}. This record does not assert independent reproduction.

### RadGrounder

Closed Accuracy: 64.5. Result ID: 6a350275fbc373592c32841e.

[Reporting paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

Recorded evaluation configuration: {"Training Data":"Report, VQA, Slake, VQA-RAD, Detect.","Vision Encoder":"Unfreeze orig. SL."}. This record does not assert independent reproduction.

### RadGrounder

Closed Accuracy: 63.2. Result ID: 6a350275fbc373592c328422.

[Reporting paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

Recorded evaluation configuration: {"Training Data":"Report, VQA, Slake, VQA-RAD"}. This record does not assert independent reproduction.

### RadGrounder

Closed Accuracy: 62.7. Result ID: 6a350275fbc373592c32841d.

[Reporting paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

Recorded evaluation configuration: {"Training Data":"Report, VQA, Slake, VQA-RAD, Detect.","Vision Encoder":"Unfreeze finet. SL."}. This record does not assert independent reproduction.

### PaliG. 2 (3B)

Closed Accuracy: 55.6. Result ID: 6a350275fbc373592c328418.

[Reporting paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76001e209131cea6ad66.md>)

Recorded evaluation configuration: {"Parameters":"3B"}. This record does not assert independent reproduction.

### RadGrounder

Closed Accuracy: 52.9. Result ID: 6a350275fbc373592c328420.

[Reporting paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

Recorded evaluation configuration: {"Training Data":"Report, VQA"}. This record does not assert independent reproduction.

### RadGrounder

Closed Accuracy: 52.9. Result ID: 6a350275fbc373592c328423.

[Reporting paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

Recorded evaluation configuration: {"Training Data":"Report, VQA, Segment"}. This record does not assert independent reproduction.

### Gemma 3 (4B)

Closed Accuracy: 33.6. Result ID: 6a350275fbc373592c328419.

[Reporting paper](<https://www.sota2.com/research/paper/6a34cd57206542251d7c61fa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa779f1e209131cea746cd.md>)

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

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