# Video Understanding on Video-Native Challenges Video-Oasis

Source: [Video Understanding on Video-Native Challenges Video-Oasis](<https://www.sota2.com/research/sota/video-understanding-on-video-native-challenges-video-oasis>)

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

Data updated: 2026-04-01T08:26:14.421Z

Evaluation of general video understanding capabilities on the Video-Native Challenges Video-Oasis dataset, assessed across fine perception, spatial world modeling, temporal dynamics, causal reasoning, and global narration.

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 Understanding |
| Dataset (including any stated split) | Video-Native Challenges Video-Oasis |
| Metric | Fine Perception |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 13 of 13 |
| 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

- [Fine Perception](<https://www.sota2.com/research/sota/video-understanding-on-video-native-challenges-video-oasis.md?metric=Fine%20Perception>): Higher is better; unit: PERCENT.

- [Spatial World Modeling](<https://www.sota2.com/research/sota/video-understanding-on-video-native-challenges-video-oasis.md?metric=Spatial%20World%20Modeling>): Higher is better; unit: PERCENT.

- [Temporal Dynamics](<https://www.sota2.com/research/sota/video-understanding-on-video-native-challenges-video-oasis.md?metric=Temporal%20Dynamics>): Higher is better; unit: PERCENT.

- [Causal Reasoning](<https://www.sota2.com/research/sota/video-understanding-on-video-native-challenges-video-oasis.md?metric=Causal%20Reasoning>): Higher is better; unit: PERCENT.

- [Global Narration](<https://www.sota2.com/research/sota/video-understanding-on-video-native-challenges-video-oasis.md?metric=Global%20Narration>): Higher is better; unit: PERCENT.

- [Overall Score](<https://www.sota2.com/research/sota/video-understanding-on-video-native-challenges-video-oasis.md?metric=Overall%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Fine Perception | Result date | Report paper |
| --- | --- | --- | --- |
| Gemini-2.5 Pro | 40.2 | Not recorded | 69cca6cc0636e5bf4afcbe64 |
| VideoTool | 31.6 | Not recorded | 69cca6cc0636e5bf4afcbe64 |
| InternVL-3.5 | 29.5 | Not recorded | 69cca6cc0636e5bf4afcbe64 |
| Qwen3-VLthink. | 29 | Not recorded | 69cca6cc0636e5bf4afcbe64 |
| VideoTree | 28.6 | Not recorded | 69cca6cc0636e5bf4afcbe64 |
| LongViLA-R1 | 28.4 | Not recorded | 69cca6cc0636e5bf4afcbe64 |
| VideoAuto-R1 | 27.5 | Not recorded | 69cca6cc0636e5bf4afcbe64 |
| Qwen3-VLinst. | 27 | Not recorded | 69cca6cc0636e5bf4afcbe64 |
| InternVL-3 | 27 | Not recorded | 69cca6cc0636e5bf4afcbe64 |
| Eagle2.5 | 26.9 | Not recorded | 69cca6cc0636e5bf4afcbe64 |
| GPT-4o | 25.6 | Not recorded | 69cca6cc0636e5bf4afcbe64 |
| Video-R1 | 24 | Not recorded | 69cca6cc0636e5bf4afcbe64 |
| Qwen2.5-VL | 23.3 | Not recorded | 69cca6cc0636e5bf4afcbe64 |

### Gemini-2.5 Pro

Fine Perception: 40.2. Result ID: 69ccd69747f34ce924b67d85.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c311e209131cea8cef5.md>)

Recorded evaluation configuration: {"Model Category":"Proprietary LLMs"}. This record does not assert independent reproduction.

### VideoTool

Fine Perception: 31.6. Result ID: 69ccd69747f34ce924b67d90.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/693ba19e6cfd488761a553d2.md>)

[Implementation](<https://github.com/fansunqi/VideoTool>)

Recorded evaluation configuration: {"Model Category":"Agentic Methods","Reasoning Model":"GPT-5-mini","Distilled evaluation setting":"true"}. This record does not assert independent reproduction.

### InternVL-3.5

Fine Perception: 29.5. Result ID: 69ccd69747f34ce924b67d8b.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7af01e209131cea85dc9.md>)

Recorded evaluation configuration: {"Model Category":"Open Source Video-LLMs","Parameter Scale":"8B"}. This record does not assert independent reproduction.

### Qwen3-VLthink.

Fine Perception: 29. Result ID: 69ccd69747f34ce924b67d88.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Model Category":"Open Source Video-LLMs","Parameter Scale":"8B"}. This record does not assert independent reproduction.

### VideoTree

Fine Perception: 28.6. Result ID: 69ccd69747f34ce924b67d8f.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa777a1e209131cea73736.md>)

[Implementation](<https://videotree2024.github.io/>)

Recorded evaluation configuration: {"Model Category":"Agentic Methods","Reasoning Model":"GPT-5-mini","Distilled evaluation setting":"true"}. This record does not assert independent reproduction.

### LongViLA-R1

Fine Perception: 28.4. Result ID: 69ccd69747f34ce924b67d8d.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a634.md>)

Recorded evaluation configuration: {"Model Category":"Open Source Video-LLMs","Parameter Scale":"7B"}. This record does not assert independent reproduction.

### VideoAuto-R1

Fine Perception: 27.5. Result ID: 69ccd69747f34ce924b67d8e.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69608b7853fdd5c36174444d.md>)

[Implementation](<https://github.com/IVUL-KAUST/VideoAuto-R1/>)

Recorded evaluation configuration: {"Model Category":"Open Source Video-LLMs","Parameter Scale":"8B"}. This record does not assert independent reproduction.

### Qwen3-VLinst.

Fine Perception: 27. Result ID: 69ccd69747f34ce924b67d87.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Model Category":"Open Source Video-LLMs","Parameter Scale":"8B"}. This record does not assert independent reproduction.

### InternVL-3

Fine Perception: 27. Result ID: 69ccd69747f34ce924b67d8a.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78241e209131cea76fa5.md>)

Recorded evaluation configuration: {"Model Category":"Open Source Video-LLMs","Parameter Scale":"8B"}. This record does not assert independent reproduction.

### Eagle2.5

Fine Perception: 26.9. Result ID: 69ccd69747f34ce924b67d89.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78331e209131cea7725b.md>)

Recorded evaluation configuration: {"Model Category":"Open Source Video-LLMs","Parameter Scale":"8B"}. This record does not assert independent reproduction.

### GPT-4o

Fine Perception: 25.6. Result ID: 69ccd69747f34ce924b67d84.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75681e209131cea67a98.md>)

Recorded evaluation configuration: {"Model Category":"Proprietary LLMs"}. This record does not assert independent reproduction.

### Video-R1

Fine Perception: 24. Result ID: 69ccd69747f34ce924b67d8c.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

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

Recorded evaluation configuration: {"Model Category":"Open Source Video-LLMs","Parameter Scale":"7B"}. This record does not assert independent reproduction.

### Qwen2.5-VL

Fine Perception: 23.3. Result ID: 69ccd69747f34ce924b67d86.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6cc0636e5bf4afcbe64.md>)

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

Recorded evaluation configuration: {"Model Category":"Open Source Video-LLMs","Parameter Scale":"7B"}. This record does not assert independent reproduction.

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