# Video Question Answering on VideoInstruct

Source: [Video Question Answering on VideoInstruct](<https://www.sota2.com/research/sota/video-question-answering-on-videoinstruct>)

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

Data updated: 2026-03-10T08:20:53.940Z

Evaluation of model performance on the VideoInstruct dataset for Video Question Answering tasks, reporting metrics related to temporal understanding and detail orientation.

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) | VideoInstruct |
| Metric | Temporal Understanding |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 2 of 2 |
| 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

- [Temporal Understanding](<https://www.sota2.com/research/sota/video-question-answering-on-videoinstruct.md?metric=Temporal%20Understanding>): Higher is better; unit: SCALAR.

- [Detail Orientation](<https://www.sota2.com/research/sota/video-question-answering-on-videoinstruct.md?metric=Detail%20Orientation>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Temporal Understanding | Result date | Report paper |
| --- | --- | --- | --- |
| T2SGrid | 2.52 | Not recorded | 69afa5642d603e6de3e43841 |
| Qwen2-VL-7B | 2.47 | Not recorded | 69afa5642d603e6de3e43841 |

### T2SGrid

Temporal Understanding: 2.52. Result ID: 69afd4636dbb7075b9581db0.

[Reporting paper](<https://www.sota2.com/research/paper/69afa5642d603e6de3e43841.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69afa5642d603e6de3e43841.md>)

Recorded evaluation configuration: {"Base Model":"Qwen2-VL-7B"}. This record does not assert independent reproduction.

### Qwen2-VL-7B

Temporal Understanding: 2.47. Result ID: 69afd4636dbb7075b9581daf.

[Reporting paper](<https://www.sota2.com/research/paper/69afa5642d603e6de3e43841.md>)

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

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

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

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