# Video Classification on MMEB Video Classification (Kinetics-700, SSv2, HMDB, UCF, Breakfast) v2 (test)

Source: [Video Classification on MMEB Video Classification (Kinetics-700, SSv2, HMDB, UCF, Breakfast) v2 (test)](<https://www.sota2.com/research/sota/video-classification-on-mmeb-video-classification-kinetics-700-ssv2-hmdb-ucf-breakfast-v2-test>)

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

Data updated: 2026-02-27T11:42:08.242Z

Evaluation of Video Classification performance on the test split of the MMEB V2 suite, aggregating results from Kinetics-700, SSv2, HMDB, UCF, and Breakfast datasets.

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 Classification |
| Dataset (including any stated split) | MMEB Video Classification (Kinetics-700, SSv2, HMDB, UCF, Breakfast) v2 (test) |
| Metric | Classification Accuracy |
| Unit | PERCENT |
| 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

- [Classification Accuracy](<https://www.sota2.com/research/sota/video-classification-on-mmeb-video-classification-kinetics-700-ssv2-hmdb-ucf-breakfast-v2-test.md?metric=Classification%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Classification Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| TARA (Tarsier-7B) | 63.7 | Not recorded | 6940e7de456146936284fc45 |
| FreeRet (Qwen2VL-7B) | 63.2 | Not recorded | 6940e7de456146936284fc45 |
| Think-Then-Embed-7B | 57.5 | Not recorded | 6940e7de456146936284fc45 |
| VLM2Vec-V2 (7B) | 45.9 | Not recorded | 6940e7de456146936284fc45 |
| VLM2Vec-Qwen2VL (7B) | 39.1 | Not recorded | 6940e7de456146936284fc45 |
| GME (7B) | 37.4 | Not recorded | 6940e7de456146936284fc45 |
| LamRA-Qwen2.5 (7B) | 32.9 | Not recorded | 6940e7de456146936284fc45 |
| ColPali v1.3 (3B) | 26.7 | Not recorded | 6940e7de456146936284fc45 |

### TARA (Tarsier-7B)

Classification Accuracy: 63.7. Result ID: 69a1830035dfcabb00e7851c.

[Reporting paper](<https://www.sota2.com/research/paper/6940e7de456146936284fc45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6940e7de456146936284fc45.md>)

Recorded evaluation configuration: {"Backbone":"Tarsier-7B","Zero-shot":"true"}. This record does not assert independent reproduction.

### FreeRet (Qwen2VL-7B)

Classification Accuracy: 63.2. Result ID: 69a1830035dfcabb00e7851b.

[Reporting paper](<https://www.sota2.com/research/paper/6940e7de456146936284fc45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bb81e209131cea8a223.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2VL-7B","Zero-shot":"true"}. This record does not assert independent reproduction.

### Think-Then-Embed-7B

Classification Accuracy: 57.5. Result ID: 69a1830035dfcabb00e7851a.

[Reporting paper](<https://www.sota2.com/research/paper/6940e7de456146936284fc45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c37d.md>)

Recorded evaluation configuration: {"Number of Parameters":"7B","Zero-shot":"true"}. This record does not assert independent reproduction.

### VLM2Vec-V2 (7B)

Classification Accuracy: 45.9. Result ID: 69a1830035dfcabb00e78519.

[Reporting paper](<https://www.sota2.com/research/paper/6940e7de456146936284fc45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79d61e209131cea7fa13.md>)

Recorded evaluation configuration: {"Number of Parameters":"7B","Zero-shot":"true"}. This record does not assert independent reproduction.

### VLM2Vec-Qwen2VL (7B)

Classification Accuracy: 39.1. Result ID: 69a1830035dfcabb00e78518.

[Reporting paper](<https://www.sota2.com/research/paper/6940e7de456146936284fc45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa766b1e209131cea6cd6f.md>)

Recorded evaluation configuration: {"Number of Parameters":"7B","Zero-shot":"true"}. This record does not assert independent reproduction.

### GME (7B)

Classification Accuracy: 37.4. Result ID: 69a1830035dfcabb00e78516.

[Reporting paper](<https://www.sota2.com/research/paper/6940e7de456146936284fc45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77b01e209131cea74c2c.md>)

Recorded evaluation configuration: {"Number of Parameters":"7B","Zero-shot":"true"}. This record does not assert independent reproduction.

### LamRA-Qwen2.5 (7B)

Classification Accuracy: 32.9. Result ID: 69a1830035dfcabb00e78517.

[Reporting paper](<https://www.sota2.com/research/paper/6940e7de456146936284fc45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76081e209131cea6b06a.md>)

[Implementation](<https://code-kunkun.github.io/LamRA/>)

Recorded evaluation configuration: {"Number of Parameters":"7B","Zero-shot":"true"}. This record does not assert independent reproduction.

### ColPali v1.3 (3B)

Classification Accuracy: 26.7. Result ID: 69a1830035dfcabb00e78515.

[Reporting paper](<https://www.sota2.com/research/paper/6940e7de456146936284fc45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77261e209131cea7161f.md>)

Recorded evaluation configuration: {"Number of Parameters":"3B","Zero-shot":"true"}. This record does not assert independent reproduction.

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