# Multi-modal Image Understanding on LLaVA-Next-7B Evaluation Suite

Source: [Multi-modal Image Understanding on LLaVA-Next-7B Evaluation Suite](<https://www.sota2.com/research/sota/multi-modal-image-understanding-on-llava-next-7b-evaluation-suite>)

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

Data updated: 2026-06-24T07:54:44.122Z

Evaluation of multimodal image understanding capabilities using the LLaVA-Next-7B Evaluation Suite, reporting scores across specific sub-benchmarks including GQA, MMB, MME, POPE, SQA, VQAText, VizWiz, and OCRBench.

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 Image Understanding |
| Dataset (including any stated split) | LLaVA-Next-7B Evaluation Suite |
| Metric | GQA Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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/multi-modal-image-understanding-on-llava-next-7b-evaluation-suite.md?metric=GQA%20Score>): Higher is better; unit: PERCENT.

- [MMB Score](<https://www.sota2.com/research/sota/multi-modal-image-understanding-on-llava-next-7b-evaluation-suite.md?metric=MMB%20Score>): Higher is better; unit: PERCENT.

- [MMB (CN) Score](<https://www.sota2.com/research/sota/multi-modal-image-understanding-on-llava-next-7b-evaluation-suite.md?metric=MMB%20(CN)%20Score>): Higher is better; unit: PERCENT.

- [MME Score](<https://www.sota2.com/research/sota/multi-modal-image-understanding-on-llava-next-7b-evaluation-suite.md?metric=MME%20Score>): Higher is better; unit: SCALAR.

- [POPE Score](<https://www.sota2.com/research/sota/multi-modal-image-understanding-on-llava-next-7b-evaluation-suite.md?metric=POPE%20Score>): Higher is better; unit: PERCENT.

- [SQA Score](<https://www.sota2.com/research/sota/multi-modal-image-understanding-on-llava-next-7b-evaluation-suite.md?metric=SQA%20Score>): Higher is better; unit: PERCENT.

- [VQAText Score](<https://www.sota2.com/research/sota/multi-modal-image-understanding-on-llava-next-7b-evaluation-suite.md?metric=VQAText%20Score>): Higher is better; unit: PERCENT.

- [VizWiz Score](<https://www.sota2.com/research/sota/multi-modal-image-understanding-on-llava-next-7b-evaluation-suite.md?metric=VizWiz%20Score>): Higher is better; unit: PERCENT.

- [OCRBench Score](<https://www.sota2.com/research/sota/multi-modal-image-understanding-on-llava-next-7b-evaluation-suite.md?metric=OCRBench%20Score>): Higher is better; unit: PERCENT.

- [Average Score](<https://www.sota2.com/research/sota/multi-modal-image-understanding-on-llava-next-7b-evaluation-suite.md?metric=Average%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | GQA Score | Result date | Report paper |
| --- | --- | --- | --- |
| Vanilla | 64.2 | Not recorded | 6a3b6485206542251d7c9676 |
| CLSE | 62.6 | Not recorded | 6a3b6485206542251d7c9676 |
| HiRED | 59.3 | Not recorded | 6a3b6485206542251d7c9676 |
| PDrop | 56.4 | Not recorded | 6a3b6485206542251d7c9676 |
| SparseVLM | 56.1 | Not recorded | 6a3b6485206542251d7c9676 |
| FastV | 55.9 | Not recorded | 6a3b6485206542251d7c9676 |
| PruMerge | 53.6 | Not recorded | 6a3b6485206542251d7c9676 |

### Vanilla

GQA Score: 64.2. Result ID: 6a3b8d3ffbc373592c4404b2.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b6485206542251d7c9676.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75151e209131cea65976.md>)

[Implementation](<https://github.com/QwenLM/Qwen2-VL>)

Recorded evaluation configuration: {"Token Budget":"2880 Tokens","Token Reduction Ratio":"100%"}. This record does not assert independent reproduction.

### CLSE

GQA Score: 62.6. Result ID: 6a3b8d3ffbc373592c4404b8.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b6485206542251d7c9676.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3b6485206542251d7c9676.md>)

Recorded evaluation configuration: {"Token Budget":"320 Tokens","Token Reduction Ratio":"88.9%"}. This record does not assert independent reproduction.

### HiRED

GQA Score: 59.3. Result ID: 6a3b8d3ffbc373592c4404b4.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b6485206542251d7c9676.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76581e209131cea6c829.md>)

Recorded evaluation configuration: {"Token Budget":"320 Tokens","Token Reduction Ratio":"88.9%"}. This record does not assert independent reproduction.

### PDrop

GQA Score: 56.4. Result ID: 6a3b8d3ffbc373592c4404b7.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b6485206542251d7c9676.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa772c1e209131cea71934.md>)

Recorded evaluation configuration: {"Token Budget":"320 Tokens","Token Reduction Ratio":"88.9%"}. This record does not assert independent reproduction.

### SparseVLM

GQA Score: 56.1. Result ID: 6a3b8d3ffbc373592c4404b6.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b6485206542251d7c9676.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79181e209131cea7c1b1.md>)

Recorded evaluation configuration: {"Token Budget":"320 Tokens","Token Reduction Ratio":"88.9%"}. This record does not assert independent reproduction.

### FastV

GQA Score: 55.9. Result ID: 6a3b8d3ffbc373592c4404b3.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b6485206542251d7c9676.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b7.md>)

Recorded evaluation configuration: {"Token Budget":"320 Tokens","Token Reduction Ratio":"88.9%"}. This record does not assert independent reproduction.

### PruMerge

GQA Score: 53.6. Result ID: 6a3b8d3ffbc373592c4404b5.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b6485206542251d7c9676.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73411e209131cea5b51f.md>)

Recorded evaluation configuration: {"Token Budget":"320 Tokens","Token Reduction Ratio":"88.9%"}. This record does not assert independent reproduction.

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