# Multi-image Visual Question Answering on MMXU (test)

Source: [Multi-image Visual Question Answering on MMXU (test)](<https://www.sota2.com/research/sota/multi-image-visual-question-answering-on-mmxu-test>)

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

Data updated: 2026-02-26T06:09:01.272Z

Evaluation of model performance on the test split of the MMXU dataset for multi-image visual question answering capabilities.

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-image Visual Question Answering |
| Dataset (including any stated split) | MMXU (test) |
| Metric | Worsen Rate |
| Unit | PERCENT |
| Direction | Lower 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

- [Worsen Rate](<https://www.sota2.com/research/sota/multi-image-visual-question-answering-on-mmxu-test.md?metric=Worsen%20Rate>): Lower is better; unit: PERCENT.

- [Improved Rate](<https://www.sota2.com/research/sota/multi-image-visual-question-answering-on-mmxu-test.md?metric=Improved%20Rate>): Higher is better; unit: PERCENT.

- [No Change Rate](<https://www.sota2.com/research/sota/multi-image-visual-question-answering-on-mmxu-test.md?metric=No%20Change%20Rate>): Higher is better; unit: PERCENT.

- [Overall Score](<https://www.sota2.com/research/sota/multi-image-visual-question-answering-on-mmxu-test.md?metric=Overall%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Worsen Rate | Result date | Report paper |
| --- | --- | --- | --- |
| Med-Flamingo | 28.9 | Not recorded | 68fa78b41e209131cea79e2c |
| Llava-med | 29.3 | Not recorded | 68fa78b41e209131cea79e2c |
| HuatuoGPT-Vision | 62.9 | Not recorded | 68fa78b41e209131cea79e2c |

### Med-Flamingo

Worsen Rate: 28.9. Result ID: 69959c3f3badfaffe6981600.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78b41e209131cea79e2c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70221e209131cea49487.md>)

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

### Llava-med

Worsen Rate: 29.3. Result ID: 69959c3f3badfaffe6981601.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78b41e209131cea79e2c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f8e1e209131cea46356.md>)

[Implementation](<https://aka.ms/llava-med>)

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

### HuatuoGPT-Vision

Worsen Rate: 62.9. Result ID: 69959c3f3badfaffe6981602.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78b41e209131cea79e2c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74f71e209131cea64ea0.md>)

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

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