# Multi-discipline Long Video Understanding on MLVU (Context Length)

Source: [Multi-discipline Long Video Understanding on MLVU (Context Length)](<https://www.sota2.com/research/sota/multi-discipline-long-video-understanding-on-mlvu-context-length>)

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

Data updated: 2026-05-27T08:52:40.731Z

Evaluation of model performance on the MLVU dataset for Multi-discipline Long Video Understanding tasks, specifically reporting the average single-forward context length achieved.

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-discipline Long Video Understanding |
| Dataset (including any stated split) | MLVU |
| Metric | Context Length (tokens) |
| Unit | SIZE |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

- [Context Length (tokens)](<https://www.sota2.com/research/sota/multi-discipline-long-video-understanding-on-mlvu-context-length.md?metric=Context%20Length%20(tokens)>): Higher is better; unit: SIZE.

## Recorded results

| Method | Context Length (tokens) | Result date | Report paper |
| --- | --- | --- | --- |
| Video-o3 | 61140 | Not recorded | 6a167a7fdbd7150506f7ff1a |
| VideoChat-R1.5-7B | 49872.9 | Not recorded | 6a167a7fdbd7150506f7ff1a |
| LongVT-RFT | 49606.9 | Not recorded | 6a167a7fdbd7150506f7ff1a |
| VideoZoomer | 17518.3 | Not recorded | 6a167a7fdbd7150506f7ff1a |
| LOVE-R1-stage3 | 16919.5 | Not recorded | 6a167a7fdbd7150506f7ff1a |
| DynFrame-8B | 7964.2 | Not recorded | 6a167a7fdbd7150506f7ff1a |
| Video-R1-7B | 3968.4 | Not recorded | 6a167a7fdbd7150506f7ff1a |
| Qwen2.5-VL-7B | 1991.5 | Not recorded | 6a167a7fdbd7150506f7ff1a |

### Video-o3

Context Length (tokens): 61140. Result ID: 6a16b0ccfbc373592c0ff87c.

[Reporting paper](<https://www.sota2.com/research/paper/6a167a7fdbd7150506f7ff1a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69802fdda8b01e76cc8e185e.md>)

Recorded evaluation configuration: {"Max retrieval / injection":"8"}. This record does not assert independent reproduction.

### VideoChat-R1.5-7B

Context Length (tokens): 49872.9. Result ID: 6a16b0ccfbc373592c0ff87d.

[Reporting paper](<https://www.sota2.com/research/paper/6a167a7fdbd7150506f7ff1a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b8f1e209131cea891be.md>)

Recorded evaluation configuration: {"Max retrieval / injection":"2"}. This record does not assert independent reproduction.

### LongVT-RFT

Context Length (tokens): 49606.9. Result ID: 6a16b0ccfbc373592c0ff87e.

[Reporting paper](<https://www.sota2.com/research/paper/6a167a7fdbd7150506f7ff1a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db35d6b0dd3342930e0c.md>)

[Implementation](<https://github.com/EvolvingLMMs-Lab/LongVT>)

Recorded evaluation configuration: {"Max retrieval / injection":"5"}. This record does not assert independent reproduction.

### VideoZoomer

Context Length (tokens): 17518.3. Result ID: 6a16b0ccfbc373592c0ff87b.

[Reporting paper](<https://www.sota2.com/research/paper/6a167a7fdbd7150506f7ff1a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69535c7e53fdd5c36173d88e.md>)

Recorded evaluation configuration: {"Max retrieval / injection":"4"}. This record does not assert independent reproduction.

### LOVE-R1-stage3

Context Length (tokens): 16919.5. Result ID: 6a16b0ccfbc373592c0ff87f.

[Reporting paper](<https://www.sota2.com/research/paper/6a167a7fdbd7150506f7ff1a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bae1e209131cea89d48.md>)

Recorded evaluation configuration: {"Max retrieval / injection":"2"}. This record does not assert independent reproduction.

### DynFrame-8B

Context Length (tokens): 7964.2. Result ID: 6a16b0ccfbc373592c0ff880.

[Reporting paper](<https://www.sota2.com/research/paper/6a167a7fdbd7150506f7ff1a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a167a7fdbd7150506f7ff1a.md>)

Recorded evaluation configuration: {"Max retrieval / injection":"1"}. This record does not assert independent reproduction.

### Video-R1-7B

Context Length (tokens): 3968.4. Result ID: 6a16b0ccfbc373592c0ff87a.

[Reporting paper](<https://www.sota2.com/research/paper/6a167a7fdbd7150506f7ff1a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c6f1e209131cea8e0ec.md>)

Recorded evaluation configuration: {"Max retrieval / injection":"0"}. This record does not assert independent reproduction.

### Qwen2.5-VL-7B

Context Length (tokens): 1991.5. Result ID: 6a16b0ccfbc373592c0ff879.

[Reporting paper](<https://www.sota2.com/research/paper/6a167a7fdbd7150506f7ff1a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa773a1e209131cea71d3c.md>)

Recorded evaluation configuration: {"Max retrieval / injection":"0"}. This record does not assert independent reproduction.

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