# Long Video Understanding on LongVideoBench (LongVB)

Source: [Long Video Understanding on LongVideoBench (LongVB)](<https://www.sota2.com/research/sota/long-video-understanding-on-longvideobench-longvb>)

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

Data updated: 2026-02-26T05:30:31.224Z

Evaluation of Long Video Understanding capabilities using the LongVideoBench (LongVB) dataset, measured by Accuracy.

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 | Long Video Understanding |
| Dataset (including any stated split) | LongVideoBench (LongVB) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 17 of 17 |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/long-video-understanding-on-longvideobench-longvb.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Gemini 2.5 Pro | 69.2 | Not recorded | 68fa78791e209131cea788a1 |
| GPT4-O | 66.7 | Not recorded | 68fa762f1e209131cea6bbab |
| Seed1.5-VL | 64.6 | Not recorded | 68fa78791e209131cea788a1 |
| Gemini 1.5 Pro | 64.4 | Not recorded | 68fa762f1e209131cea6bbab |
| Seed1.5-VL | 64 | Not recorded | 68fa78791e209131cea788a1 |
| GPT4-V | 60.7 | Not recorded | 68fa762f1e209131cea6bbab |
| LinVT-InternVL2 | 57.4 | Not recorded | 68fa762f1e209131cea6bbab |
| LinVT-Qwen2-VL | 57.2 | Not recorded | 68fa762f1e209131cea6bbab |
| LinVT-Molmo | 56.8 | Not recorded | 68fa762f1e209131cea6bbab |
| LinVT-BLIP-3 | 56.6 | Not recorded | 68fa762f1e209131cea6bbab |
| Oryx | 55.3 | Not recorded | 68fa762f1e209131cea6bbab |
| Kangaroo | 54.8 | Not recorded | 68fa762f1e209131cea6bbab |
| PLLAVA | 53.5 | Not recorded | 68fa762f1e209131cea6bbab |
| LinVT-Aquila | 53.5 | Not recorded | 68fa762f1e209131cea6bbab |
| LinVT-Mipha | 49.7 | Not recorded | 68fa762f1e209131cea6bbab |
| ShareGPT4Video | 41.8 | Not recorded | 68fa762f1e209131cea6bbab |
| Video-LLaVA | 37.6 | Not recorded | 68fa762f1e209131cea6bbab |

### Gemini 2.5 Pro

Accuracy: 69.2. Result ID: 69958c593badfaffe685f8be.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78791e209131cea788a1.md>)

Recorded evaluation configuration: {"Evaluation frame rate":"1 FPS","self-collected via API":"true"}. This record does not assert independent reproduction.

### GPT4-O

Accuracy: 66.7. Result ID: 69942ab03a6c3148491b8399.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

Recorded evaluation configuration: {"Size":null}. This record does not assert independent reproduction.

### Seed1.5-VL

Accuracy: 64.6. Result ID: 69958c593badfaffe685f8bc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78791e209131cea788a1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78791e209131cea788a1.md>)

Recorded evaluation configuration: {"Thinking mode":"true","Evaluation frame rate":"1 FPS"}. This record does not assert independent reproduction.

### Gemini 1.5 Pro

Accuracy: 64.4. Result ID: 69942ab03a6c3148491b839a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76361e209131cea6be11.md>)

Recorded evaluation configuration: {"Size":null}. This record does not assert independent reproduction.

### Seed1.5-VL

Accuracy: 64. Result ID: 69958c593badfaffe685f8bd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78791e209131cea788a1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78791e209131cea788a1.md>)

Recorded evaluation configuration: {"Thinking mode":"false","Evaluation frame rate":"1 FPS"}. This record does not assert independent reproduction.

### GPT4-V

Accuracy: 60.7. Result ID: 69942ab03a6c3148491b8398.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

Recorded evaluation configuration: {"Size":null}. This record does not assert independent reproduction.

### LinVT-InternVL2

Accuracy: 57.4. Result ID: 69942ab03a6c3148491b83a5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Implementation](<https://github.com/gls0425/LinVT>)

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

### LinVT-Qwen2-VL

Accuracy: 57.2. Result ID: 69942ab03a6c3148491b83a4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Implementation](<https://github.com/gls0425/LinVT>)

Recorded evaluation configuration: {"Size":"7B"}. This record does not assert independent reproduction.

### LinVT-Molmo

Accuracy: 56.8. Result ID: 69942ab03a6c3148491b83a3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Implementation](<https://github.com/gls0425/LinVT>)

Recorded evaluation configuration: {"Size":"7B"}. This record does not assert independent reproduction.

### LinVT-BLIP-3

Accuracy: 56.6. Result ID: 69942ab03a6c3148491b83a2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Implementation](<https://github.com/gls0425/LinVT>)

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

### Oryx

Accuracy: 55.3. Result ID: 69942ab03a6c3148491b839f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa772c1e209131cea71918.md>)

[Implementation](<https://github.com/Oryx-mllm/Oryx>)

Recorded evaluation configuration: {"Size":"7B"}. This record does not assert independent reproduction.

### Kangaroo

Accuracy: 54.8. Result ID: 69942ab03a6c3148491b839b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74a91e209131cea6306e.md>)

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

### PLLAVA

Accuracy: 53.5. Result ID: 69942ab03a6c3148491b839c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73101e209131cea5a3ea.md>)

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

Recorded evaluation configuration: {"Size":"34B"}. This record does not assert independent reproduction.

### LinVT-Aquila

Accuracy: 53.5. Result ID: 69942ab03a6c3148491b83a1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Implementation](<https://github.com/gls0425/LinVT>)

Recorded evaluation configuration: {"Size":"2B"}. This record does not assert independent reproduction.

### LinVT-Mipha

Accuracy: 49.7. Result ID: 69942ab03a6c3148491b83a0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Implementation](<https://github.com/gls0425/LinVT>)

Recorded evaluation configuration: {"Size":"1.6B"}. This record does not assert independent reproduction.

### ShareGPT4Video

Accuracy: 41.8. Result ID: 69942ab03a6c3148491b839d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738e1e209131cea5d310.md>)

[Implementation](<https://sharegpt4video.github.io/>)

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

### Video-LLaVA

Accuracy: 37.6. Result ID: 69942ab03a6c3148491b839e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa762f1e209131cea6bbab.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75051e209131cea65534.md>)

[Implementation](<https://github.com/PKU-YuanGroup/Video-LLaVA>)

Recorded evaluation configuration: {"Size":"7B"}. This record does not assert independent reproduction.

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