# Multi-modal Multi-span Medical Question Answering on M3QuestionIng

Source: [Multi-modal Multi-span Medical Question Answering on M3QuestionIng](<https://www.sota2.com/research/sota/multi-modal-multi-span-medical-question-answering-on-m3questioning>)

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

Data updated: 2026-06-30T06:39:46.349Z

Evaluation of model performance on the M3QuestionIng dataset for Multi-modal Multi-span Medical Question Answering, typically measured using Accuracy, F1-Score, and Exact Match.

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 | Multi-modal Multi-span Medical Question Answering |
| Dataset (including any stated split) | M3QuestionIng |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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/multi-modal-multi-span-medical-question-answering-on-m3questioning.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [F1-Score](<https://www.sota2.com/research/sota/multi-modal-multi-span-medical-question-answering-on-m3questioning.md?metric=F1-Score>): Higher is better; unit: PERCENT.

- [Exact Match](<https://www.sota2.com/research/sota/multi-modal-multi-span-medical-question-answering-on-m3questioning.md?metric=Exact%20Match>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| M3QAFrame | 94.34 | Not recorded | 6a434d6eba73165dbb7437c5 |
| GPT-4o | 81.2 | Not recorded | 6a434d6eba73165dbb7437c5 |
| Qwen2.5-VL | 65.38 | Not recorded | 6a434d6eba73165dbb7437c5 |
| MMEmbed | 52.15 | Not recorded | 6a434d6eba73165dbb7437c5 |
| LLaVA | 40.56 | Not recorded | 6a434d6eba73165dbb7437c5 |
| MedGemma | 39.16 | Not recorded | 6a434d6eba73165dbb7437c5 |
| VLM2Vec | 28.05 | Not recorded | 6a434d6eba73165dbb7437c5 |
| Uni-MedCLIP | 12.74 | Not recorded | 6a434d6eba73165dbb7437c5 |

### M3QAFrame

Accuracy: 94.34. Result ID: 6a436497fbc373592c92bdeb.

[Reporting paper](<https://www.sota2.com/research/paper/6a434d6eba73165dbb7437c5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a434d6eba73165dbb7437c5.md>)

Recorded evaluation configuration: {"Training Strategy":"Trained on M3QuestionIng"}. This record does not assert independent reproduction.

### GPT-4o

Accuracy: 81.2. Result ID: 6a436497fbc373592c92bdea.

[Reporting paper](<https://www.sota2.com/research/paper/6a434d6eba73165dbb7437c5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75681e209131cea67a98.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"Average of zero-shot and few-shot"}. This record does not assert independent reproduction.

### Qwen2.5-VL

Accuracy: 65.38. Result ID: 6a436497fbc373592c92bde8.

[Reporting paper](<https://www.sota2.com/research/paper/6a434d6eba73165dbb7437c5.md>)

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

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

### MMEmbed

Accuracy: 52.15. Result ID: 6a436497fbc373592c92bde5.

[Reporting paper](<https://www.sota2.com/research/paper/6a434d6eba73165dbb7437c5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77201e209131cea713da.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"Average of zero-shot and few-shot"}. This record does not assert independent reproduction.

### LLaVA

Accuracy: 40.56. Result ID: 6a436497fbc373592c92bde7.

[Reporting paper](<https://www.sota2.com/research/paper/6a434d6eba73165dbb7437c5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa717d1e209131cea51970.md>)

[Implementation](<https://llava-vl.github.io/>)

Recorded evaluation configuration: {"Evaluation Protocol":"Average of zero-shot and few-shot"}. This record does not assert independent reproduction.

### MedGemma

Accuracy: 39.16. Result ID: 6a436497fbc373592c92bde9.

[Reporting paper](<https://www.sota2.com/research/paper/6a434d6eba73165dbb7437c5.md>)

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

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

### VLM2Vec

Accuracy: 28.05. Result ID: 6a436497fbc373592c92bde6.

[Reporting paper](<https://www.sota2.com/research/paper/6a434d6eba73165dbb7437c5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa766b1e209131cea6cd6f.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"Average of zero-shot and few-shot"}. This record does not assert independent reproduction.

### Uni-MedCLIP

Accuracy: 12.74. Result ID: 6a436497fbc373592c92bde4.

[Reporting paper](<https://www.sota2.com/research/paper/6a434d6eba73165dbb7437c5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa763c1e209131cea6c048.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"Average of zero-shot and few-shot"}. This record does not assert independent reproduction.

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