# Video Question Answering on EventBench 60-1800s (test)

Source: [Video Question Answering on EventBench 60-1800s (test)](<https://www.sota2.com/research/sota/video-question-answering-on-eventbench-60-1800s-test>)

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

Data updated: 2026-05-08T10:21:04.590Z

Evaluation of model performance on the test split of EventBench (60-1800s duration videos) for Video Question Answering, reporting metrics across Contextual, Episodic, and Counter-intuitive question types.

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) | EventBench 60-1800s (test) |
| Metric | Contextual 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

- [Contextual Accuracy](<https://www.sota2.com/research/sota/video-question-answering-on-eventbench-60-1800s-test.md?metric=Contextual%20Accuracy>): Higher is better; unit: PERCENT.

- [Episodic Accuracy](<https://www.sota2.com/research/sota/video-question-answering-on-eventbench-60-1800s-test.md?metric=Episodic%20Accuracy>): Higher is better; unit: PERCENT.

- [Counter-intuitive Accuracy](<https://www.sota2.com/research/sota/video-question-answering-on-eventbench-60-1800s-test.md?metric=Counter-intuitive%20Accuracy>): Higher is better; unit: PERCENT.

- [Overall Accuracy](<https://www.sota2.com/research/sota/video-question-answering-on-eventbench-60-1800s-test.md?metric=Overall%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Contextual Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| InternVL3.5-8B + EC-RAG | 62.28 | Not recorded | 69fd6e1a5497b3b670c8946a |
| Qwen3-VL-8B + EC-RAG | 61.77 | Not recorded | 69fd6e1a5497b3b670c8946a |
| VideoLlama3-7B + EC-RAG | 61.01 | Not recorded | 69fd6e1a5497b3b670c8946a |
| InternVL3.5-8B | 56.96 | Not recorded | 69fd6e1a5497b3b670c8946a |
| VideoLlama3-7B | 56.71 | Not recorded | 69fd6e1a5497b3b670c8946a |
| Qwen3-VL-8B | 53.16 | Not recorded | 69fd6e1a5497b3b670c8946a |
| GPT-4o (Closed-source) | 50.13 | Not recorded | 69fd6e1a5497b3b670c8946a |
| Gemini1.5-pro (Closed-source) | 32.15 | Not recorded | 69fd6e1a5497b3b670c8946a |

### InternVL3.5-8B + EC-RAG

Contextual Accuracy: 62.28. Result ID: 69fdb90847f34ce924e14890.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e1a5497b3b670c8946a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69fd6e1a5497b3b670c8946a.md>)

Recorded evaluation configuration: {"Backbone":"InternVL3.5-8B","Augmentation":"EC-RAG"}. This record does not assert independent reproduction.

### Qwen3-VL-8B + EC-RAG

Contextual Accuracy: 61.77. Result ID: 69fdb90847f34ce924e1488c.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e1a5497b3b670c8946a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69fd6e1a5497b3b670c8946a.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B","Augmentation":"EC-RAG"}. This record does not assert independent reproduction.

### VideoLlama3-7B + EC-RAG

Contextual Accuracy: 61.01. Result ID: 69fdb90847f34ce924e1488e.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e1a5497b3b670c8946a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69fd6e1a5497b3b670c8946a.md>)

Recorded evaluation configuration: {"Backbone":"VideoLlama3-7B","Augmentation":"EC-RAG"}. This record does not assert independent reproduction.

### InternVL3.5-8B

Contextual Accuracy: 56.96. Result ID: 69fdb90847f34ce924e1488f.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e1a5497b3b670c8946a.md>)

Recorded evaluation configuration: {"Backbone":"InternVL3.5-8B"}. This record does not assert independent reproduction.

### VideoLlama3-7B

Contextual Accuracy: 56.71. Result ID: 69fdb90847f34ce924e1488d.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e1a5497b3b670c8946a.md>)

Recorded evaluation configuration: {"Backbone":"VideoLlama3-7B"}. This record does not assert independent reproduction.

### Qwen3-VL-8B

Contextual Accuracy: 53.16. Result ID: 69fdb90847f34ce924e1488b.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e1a5497b3b670c8946a.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction.

### GPT-4o (Closed-source)

Contextual Accuracy: 50.13. Result ID: 69fdb90847f34ce924e14889.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e1a5497b3b670c8946a.md>)

Recorded evaluation configuration: {"Access":"Closed-source"}. This record does not assert independent reproduction.

### Gemini1.5-pro (Closed-source)

Contextual Accuracy: 32.15. Result ID: 69fdb90847f34ce924e1488a.

[Reporting paper](<https://www.sota2.com/research/paper/69fd6e1a5497b3b670c8946a.md>)

Recorded evaluation configuration: {"Access":"Closed-source"}. This record does not assert independent reproduction.

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