# Video Question Answering on MLVU (p@1, p@4, m@4)

Source: [Video Question Answering on MLVU (p@1, p@4, m@4)](<https://www.sota2.com/research/sota/video-question-answering-on-mlvu-p-at-1-p-at-4-m-at-4>)

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

Data updated: 2026-07-07T07:55:25.455Z

Evaluation of Video Question Answering performance on the MLVU dataset, measured using Precision at K (p@1, p@4) and Mean at 4 (m@4).

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 | Video Question Answering |
| Dataset (including any stated split) | MLVU |
| Metric | p@1 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 11 of 11 |
| 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

- [p@1](<https://www.sota2.com/research/sota/video-question-answering-on-mlvu-p-at-1-p-at-4-m-at-4.md?metric=p%401>): Higher is better; unit: PERCENT.

- [p@4](<https://www.sota2.com/research/sota/video-question-answering-on-mlvu-p-at-1-p-at-4-m-at-4.md?metric=p%404>): Higher is better; unit: PERCENT.

- [m@4](<https://www.sota2.com/research/sota/video-question-answering-on-mlvu-p-at-1-p-at-4-m-at-4.md?metric=m%404>): Higher is better; unit: PERCENT.

## Recorded results

| Method | p@1 | Result date | Report paper |
| --- | --- | --- | --- |
| VSeek-VETL | 67.9 | Not recorded | 6a4c87dc85dba69fef8b2321 |
| VSeek-EM | 66.7 | Not recorded | 6a4c87dc85dba69fef8b2321 |
| Qwen3-VL-4B-Instruct | 61.9 | Not recorded | 6a4c87dc85dba69fef8b2321 |
| Qwen3-VL-8B-Instruct | 61.9 | Not recorded | 6a4c87dc85dba69fef8b2321 |
| GPT-5.2 | 61.3 | Not recorded | 6a4c87dc85dba69fef8b2321 |
| Qwen3-VL-4B-Thinking | 58 | Not recorded | 6a4c87dc85dba69fef8b2321 |
| VSeek-GPT | 56.3 | Not recorded | 6a4c87dc85dba69fef8b2321 |
| Qwen3-VL-8B-Thinking | 56.3 | Not recorded | 6a4c87dc85dba69fef8b2321 |
| Naive RAG | 55.6 | Not recorded | 6a4c87dc85dba69fef8b2321 |
| VideoTree | 46.1 | Not recorded | 6a4c87dc85dba69fef8b2321 |
| InternVL-3.5-4B | 41.8 | Not recorded | 6a4c87dc85dba69fef8b2321 |

### VSeek-VETL

p@1: 67.9. Result ID: 6a4cb0d260b88bb339c7f7ef.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

Recorded evaluation configuration: {"reward\_type":"VETL"}. This record does not assert independent reproduction.

### VSeek-EM

p@1: 66.7. Result ID: 6a4cb0d260b88bb339c7f7ee.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

Recorded evaluation configuration: {"reward\_type":"exact-match"}. This record does not assert independent reproduction.

### Qwen3-VL-4B-Instruct

p@1: 61.9. Result ID: 6a4cb0d260b88bb339c7f7e9.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"parameters":"4B"}. This record does not assert independent reproduction.

### Qwen3-VL-8B-Instruct

p@1: 61.9. Result ID: 6a4cb0d260b88bb339c7f7eb.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"parameters":"8B"}. This record does not assert independent reproduction.

### GPT-5.2

p@1: 61.3. Result ID: 6a4cb0d260b88bb339c7f7e5.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

Recorded evaluation configuration: {"frames":"64"}. This record does not assert independent reproduction.

### Qwen3-VL-4B-Thinking

p@1: 58. Result ID: 6a4cb0d260b88bb339c7f7e8.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"parameters":"4B"}. This record does not assert independent reproduction.

### VSeek-GPT

p@1: 56.3. Result ID: 6a4cb0d260b88bb339c7f7e6.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

Recorded evaluation configuration: {"base\_model":"GPT-5.2"}. This record does not assert independent reproduction.

### Qwen3-VL-8B-Thinking

p@1: 56.3. Result ID: 6a4cb0d260b88bb339c7f7ea.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"parameters":"8B"}. This record does not assert independent reproduction.

### Naive RAG

p@1: 55.6. Result ID: 6a4cb0d260b88bb339c7f7ec.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### VideoTree

p@1: 46.1. Result ID: 6a4cb0d260b88bb339c7f7ed.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa777a1e209131cea73736.md>)

[Implementation](<https://videotree2024.github.io/>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### InternVL-3.5-4B

p@1: 41.8. Result ID: 6a4cb0d260b88bb339c7f7e7.

[Reporting paper](<https://www.sota2.com/research/paper/6a4c87dc85dba69fef8b2321.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71de1e209131cea5370d.md>)

[Implementation](<https://github.com/OpenGVLab/InternVL>)

Recorded evaluation configuration: {"parameters":"4B"}. This record does not assert independent reproduction.

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