# Multi-turn Visual Question Answering on MT-GQA (Accuracy Metrics)

Source: [Multi-turn Visual Question Answering on MT-GQA (Accuracy Metrics)](<https://www.sota2.com/research/sota/multi-turn-visual-question-answering-on-mt-gqa-accuracy-metrics>)

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

Data updated: 2026-03-24T16:03:39.789Z

Evaluation of model performance on the MT-GQA dataset for multi-turn visual question answering, reporting accuracy metrics (Acc1, Acc2, Acc3, Avg).

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-turn Visual Question Answering |
| Dataset (including any stated split) | MT-GQA |
| Metric | Acc1 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 33 of 33 |
| 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

- [Acc1](<https://www.sota2.com/research/sota/multi-turn-visual-question-answering-on-mt-gqa-accuracy-metrics.md?metric=Acc1>): Higher is better; unit: PERCENT.

- [Acc2](<https://www.sota2.com/research/sota/multi-turn-visual-question-answering-on-mt-gqa-accuracy-metrics.md?metric=Acc2>): Higher is better; unit: PERCENT.

- [Acc3](<https://www.sota2.com/research/sota/multi-turn-visual-question-answering-on-mt-gqa-accuracy-metrics.md?metric=Acc3>): Higher is better; unit: PERCENT.

- [Avg Acc](<https://www.sota2.com/research/sota/multi-turn-visual-question-answering-on-mt-gqa-accuracy-metrics.md?metric=Avg%20Acc>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Acc1 | Result date | Report paper |
| --- | --- | --- | --- |
| Base | 65.45 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| MetaCompress | 64.86 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Sample | 63.95 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Base | 63.83 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| MetaCompress | 63.38 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Random | 62.89 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Base | 62.47 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Sample | 62.28 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| FastV | 62.1 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Random | 61.96 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Base | 61.76 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Base | 61.76 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| FastV | 61.54 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Base | 60.21 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| MetaCompress | 60.11 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| MetaCompress | 59.48 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| FastV | 58.95 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| MetaCompress | 58.62 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Sample | 58.51 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Sample | 58.48 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| FastV | 57.99 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| PruMerge | 57.89 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Random | 57.79 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Sample | 57.79 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Random | 57.74 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| PruMerge | 57.67 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Random | 56.96 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| MetaCompress | 55.95 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| VisionZip | 55.08 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| FastV | 54.79 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Sample | 54.6 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| ToMe | 53.64 | Not recorded | 69c21aa60636e5bf4afc4a72 |
| Spatial | 51.05 | Not recorded | 69c21aa60636e5bf4afc4a72 |

### Base

Acc1: 65.45. Result ID: 69c2b5c847f34ce924ac014c.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-NeXT-13b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### MetaCompress

Acc1: 64.86. Result ID: 69c2b5c847f34ce924ac0150.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-NeXT-13b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Sample

Acc1: 63.95. Result ID: 69c2b5c847f34ce924ac014e.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-NeXT-13b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Base

Acc1: 63.83. Result ID: 69c2b5c847f34ce924ac0147.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-NeXT-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### MetaCompress

Acc1: 63.38. Result ID: 69c2b5c847f34ce924ac014b.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-NeXT-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Random

Acc1: 62.89. Result ID: 69c2b5c847f34ce924ac014d.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-NeXT-13b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Base

Acc1: 62.47. Result ID: 69c2b5c847f34ce924ac0141.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-13b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Sample

Acc1: 62.28. Result ID: 69c2b5c847f34ce924ac0149.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-NeXT-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### FastV

Acc1: 62.1. Result ID: 69c2b5c847f34ce924ac014f.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a0b1e209131cea25b04.md>)

[Implementation](<https://github.com/google-research/vision_transformer>)

Recorded evaluation configuration: {"Model":"LLaVA-NeXT-13b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Random

Acc1: 61.96. Result ID: 69c2b5c847f34ce924ac0148.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-NeXT-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Base

Acc1: 61.76. Result ID: 69c2b5c847f34ce924ac013b.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Base

Acc1: 61.76. Result ID: 69c2b5c847f34ce924ac01c2.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-7b","Token Merging Method":"None (Full Tokens)","Reduction Rate":"0%"}. This record does not assert independent reproduction.

### FastV

Acc1: 61.54. Result ID: 69c2b5c847f34ce924ac014a.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a0b1e209131cea25b04.md>)

[Implementation](<https://github.com/google-research/vision_transformer>)

Recorded evaluation configuration: {"Model":"LLaVA-NeXT-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Base

Acc1: 60.21. Result ID: 69c2b5c847f34ce924ac0151.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"XComposer-2.5-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### MetaCompress

Acc1: 60.11. Result ID: 69c2b5c847f34ce924ac0155.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"XComposer-2.5-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### MetaCompress

Acc1: 59.48. Result ID: 69c2b5c847f34ce924ac0146.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-13b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### FastV

Acc1: 58.95. Result ID: 69c2b5c847f34ce924ac0154.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a0b1e209131cea25b04.md>)

[Implementation](<https://github.com/google-research/vision_transformer>)

Recorded evaluation configuration: {"Model":"XComposer-2.5-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### MetaCompress

Acc1: 58.62. Result ID: 69c2b5c847f34ce924ac0140.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Sample

Acc1: 58.51. Result ID: 69c2b5c847f34ce924ac0143.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-13b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Sample

Acc1: 58.48. Result ID: 69c2b5c847f34ce924ac013d.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### FastV

Acc1: 57.99. Result ID: 69c2b5c847f34ce924ac0144.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a0b1e209131cea25b04.md>)

[Implementation](<https://github.com/google-research/vision_transformer>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-13b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### PruMerge

Acc1: 57.89. Result ID: 69c2b5c847f34ce924ac013f.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

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

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

Recorded evaluation configuration: {"Model":"LLaVA-1.5-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Random

Acc1: 57.79. Result ID: 69c2b5c847f34ce924ac013c.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Sample

Acc1: 57.79. Result ID: 69c2b5c847f34ce924ac0153.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"XComposer-2.5-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Random

Acc1: 57.74. Result ID: 69c2b5c847f34ce924ac0142.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-13b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### PruMerge

Acc1: 57.67. Result ID: 69c2b5c847f34ce924ac0145.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

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

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

Recorded evaluation configuration: {"Model":"LLaVA-1.5-13b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Random

Acc1: 56.96. Result ID: 69c2b5c847f34ce924ac0152.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"XComposer-2.5-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### MetaCompress

Acc1: 55.95. Result ID: 69c2b5c847f34ce924ac01c7.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-7b","Token Merging Method":"Ours","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### VisionZip

Acc1: 55.08. Result ID: 69c2b5c847f34ce924ac01c6.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

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

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

Recorded evaluation configuration: {"Model":"LLaVA-1.5-7b","Token Merging Method":"VisionZip","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### FastV

Acc1: 54.79. Result ID: 69c2b5c847f34ce924ac013e.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a0b1e209131cea25b04.md>)

[Implementation](<https://github.com/google-research/vision_transformer>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-7b","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Sample

Acc1: 54.6. Result ID: 69c2b5c847f34ce924ac01c3.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-7b","Token Merging Method":"Sample","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### ToMe

Acc1: 53.64. Result ID: 69c2b5c847f34ce924ac01c5.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ea91e209131cea4130c.md>)

[Implementation](<http://github.com/facebookresearch/ToMe>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-7b","Token Merging Method":"ToMe","Reduction Rate":"70%"}. This record does not assert independent reproduction.

### Spatial

Acc1: 51.05. Result ID: 69c2b5c847f34ce924ac01c4.

[Reporting paper](<https://www.sota2.com/research/paper/69c21aa60636e5bf4afc4a72.md>)

Recorded evaluation configuration: {"Model":"LLaVA-1.5-7b","Token Merging Method":"Spatial","Reduction Rate":"70%"}. This record does not assert independent reproduction.

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