# Visual Question Answering on GQA (Score, Relative Performance %)

Source: [Visual Question Answering on GQA (Score, Relative Performance %)](<https://www.sota2.com/research/sota/visual-question-answering-on-gqa-score-relative-performance-percent>)

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

Data updated: 2026-02-27T09:53:44.077Z

Evaluation of model performance on the GQA dataset for visual question answering, reporting the primary Score and the Relative Performance percentage.

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 | Visual Question Answering |
| Dataset (including any stated split) | GQA |
| Metric | GQA Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| 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

- [GQA Score](<https://www.sota2.com/research/sota/visual-question-answering-on-gqa-score-relative-performance-percent.md?metric=GQA%20Score>): Higher is better; unit: PERCENT.

- [Relative Performance (%)](<https://www.sota2.com/research/sota/visual-question-answering-on-gqa-score-relative-performance-percent.md?metric=Relative%20Performance%20(%25)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | GQA Score | Result date | Report paper |
| --- | --- | --- | --- |
| Vanilla | 61.9 | Not recorded | 68fa760f1e209131cea6b33b |
| VisionZip | 61.5 | Not recorded | 68fa760f1e209131cea6b33b |
| VisionZip | 60 | Not recorded | 68fa760f1e209131cea6b33b |
| VisionZip | 58.9 | Not recorded | 68fa760f1e209131cea6b33b |

### Vanilla

GQA Score: 61.9. Result ID: 69a1699535dfcabb00e6e42f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa760f1e209131cea6b33b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa732e1e209131cea5af59.md>)

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

Recorded evaluation configuration: {"Base Model":"LLaVA 1.5 7B","Number of Visual Tokens":"576","Training Protocol":"Standard training"}. This record does not assert independent reproduction.

### VisionZip

GQA Score: 61.5. Result ID: 69a1699535dfcabb00e6e430.

[Reporting paper](<https://www.sota2.com/research/paper/68fa760f1e209131cea6b33b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa760f1e209131cea6b33b.md>)

[Implementation](<https://github.com/dvlab-research/VisionZip>)

Recorded evaluation configuration: {"Base Model":"LLaVA 1.5 7B","Number of Visual Tokens":"192","Training Protocol":"VisionZip training"}. This record does not assert independent reproduction.

### VisionZip

GQA Score: 60. Result ID: 69a1699535dfcabb00e6e431.

[Reporting paper](<https://www.sota2.com/research/paper/68fa760f1e209131cea6b33b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa760f1e209131cea6b33b.md>)

[Implementation](<https://github.com/dvlab-research/VisionZip>)

Recorded evaluation configuration: {"Base Model":"LLaVA 1.5 7B","Number of Visual Tokens":"128","Training Protocol":"VisionZip training"}. This record does not assert independent reproduction.

### VisionZip

GQA Score: 58.9. Result ID: 69a1699535dfcabb00e6e432.

[Reporting paper](<https://www.sota2.com/research/paper/68fa760f1e209131cea6b33b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa760f1e209131cea6b33b.md>)

[Implementation](<https://github.com/dvlab-research/VisionZip>)

Recorded evaluation configuration: {"Base Model":"LLaVA 1.5 7B","Number of Visual Tokens":"64","Training Protocol":"VisionZip training"}. This record does not assert independent reproduction.

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