# Cinematography language understanding on ShotBench 1.0 (test)

Source: [Cinematography language understanding on ShotBench 1.0 (test)](<https://www.sota2.com/research/sota/cinematography-language-understanding-on-shotbench-1-0-test>)

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

Data updated: 2026-03-31T06:59:54.516Z

Evaluation of model performance on the ShotBench 1.0 test set, specifically assessing understanding of various cinematography elements such as shot size, framing, camera angle, lighting, and composition.

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 | Cinematography language understanding |
| Dataset (including any stated split) | ShotBench 1.0 (test) |
| Metric | Shot Size Accuracy (SS) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 25 of 25 |
| 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

- [Shot Size Accuracy (SS)](<https://www.sota2.com/research/sota/cinematography-language-understanding-on-shotbench-1-0-test.md?metric=Shot%20Size%20Accuracy%20(SS)>): Higher is better; unit: PERCENT.

- [Shot Framing Accuracy (SF)](<https://www.sota2.com/research/sota/cinematography-language-understanding-on-shotbench-1-0-test.md?metric=Shot%20Framing%20Accuracy%20(SF)>): Higher is better; unit: PERCENT.

- [Camera Angle Accuracy (CA)](<https://www.sota2.com/research/sota/cinematography-language-understanding-on-shotbench-1-0-test.md?metric=Camera%20Angle%20Accuracy%20(CA)>): Higher is better; unit: PERCENT.

- [Lens Size Accuracy (LS)](<https://www.sota2.com/research/sota/cinematography-language-understanding-on-shotbench-1-0-test.md?metric=Lens%20Size%20Accuracy%20(LS)>): Higher is better; unit: PERCENT.

- [Lighting Type Accuracy (LT)](<https://www.sota2.com/research/sota/cinematography-language-understanding-on-shotbench-1-0-test.md?metric=Lighting%20Type%20Accuracy%20(LT)>): Higher is better; unit: PERCENT.

- [Lighting Conditions Accuracy (LC)](<https://www.sota2.com/research/sota/cinematography-language-understanding-on-shotbench-1-0-test.md?metric=Lighting%20Conditions%20Accuracy%20(LC)>): Higher is better; unit: PERCENT.

- [Shot Composition Accuracy (SC)](<https://www.sota2.com/research/sota/cinematography-language-understanding-on-shotbench-1-0-test.md?metric=Shot%20Composition%20Accuracy%20(SC)>): Higher is better; unit: PERCENT.

- [Camera Movement Accuracy (CM)](<https://www.sota2.com/research/sota/cinematography-language-understanding-on-shotbench-1-0-test.md?metric=Camera%20Movement%20Accuracy%20(CM)>): Higher is better; unit: PERCENT.

- [Average Accuracy](<https://www.sota2.com/research/sota/cinematography-language-understanding-on-shotbench-1-0-test.md?metric=Average%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Shot Size Accuracy (SS) | Result date | Report paper |
| --- | --- | --- | --- |
| ShotVL (3B) | 77.9 | Not recorded | 68fa79b31e209131cea7ef72 |
| Qwen2.5-VL-72B-Instruct | 75.1 | Not recorded | 68fa79b31e209131cea7ef72 |
| InternVL3-78B | 69.7 | Not recorded | 68fa79b31e209131cea7ef72 |
| GPT-4o | 69.3 | Not recorded | 68fa79b31e209131cea7ef72 |
| Qwen2.5-VL-7B-Instruct | 69.1 | Not recorded | 68fa79b31e209131cea7ef72 |
| InternVL3-38B | 68 | Not recorded | 68fa79b31e209131cea7ef72 |
| InternVL2.5-38B | 67.8 | Not recorded | 68fa79b31e209131cea7ef72 |
| Qwen2.5-VL-32B-Instruct | 62.3 | Not recorded | 68fa79b31e209131cea7ef72 |
| InternVL3-8B | 62.1 | Not recorded | 68fa79b31e209131cea7ef72 |
| InternVL3-14B | 59.6 | Not recorded | 68fa79b31e209131cea7ef72 |
| LLaVA-Onevision-Qwen2-7B-Ov-Chat | 58.4 | Not recorded | 68fa79b31e209131cea7ef72 |
| Gemini-2.5-flash-preview-04-17 | 57.7 | Not recorded | 68fa79b31e209131cea7ef72 |
| LLaVA-Video-7B-Qwen2 | 56.9 | Not recorded | 68fa79b31e209131cea7ef72 |
| InternVL2.5-8B | 56.3 | Not recorded | 68fa79b31e209131cea7ef72 |
| InternVL3-2B | 56.3 | Not recorded | 68fa79b31e209131cea7ef72 |
| Qwen2.5-VL-3B-Instruct | 54.6 | Not recorded | 68fa79b31e209131cea7ef72 |
| Internlm-xcomposer2d5-7B | 51.1 | Not recorded | 68fa79b31e209131cea7ef72 |
| Gemini-2.0-flash | 48.9 | Not recorded | 68fa79b31e209131cea7ef72 |
| VILA1.5-8B | 40.6 | Not recorded | 68fa79b31e209131cea7ef72 |
| VILA1.5-13B | 36.7 | Not recorded | 68fa79b31e209131cea7ef72 |
| LLaVA-NeXT-Video-7B | 35.9 | Not recorded | 68fa79b31e209131cea7ef72 |
| Ovis2-8B | 35.9 | Not recorded | 68fa79b31e209131cea7ef72 |
| VILA1.5-3B | 33.4 | Not recorded | 68fa79b31e209131cea7ef72 |
| Instructblip-vicuna-7B | 27 | Not recorded | 68fa79b31e209131cea7ef72 |
| Instructblip-vicuna-13B | 26.8 | Not recorded | 68fa79b31e209131cea7ef72 |

### ShotVL (3B)

Shot Size Accuracy (SS): 77.9. Result ID: 69cb6fef47f34ce924b33504.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

### Qwen2.5-VL-72B-Instruct

Shot Size Accuracy (SS): 75.1. Result ID: 69cb6fef47f34ce924b334ff.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

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

### InternVL3-78B

Shot Size Accuracy (SS): 69.7. Result ID: 69cb6fef47f34ce924b33500.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

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

### GPT-4o

Shot Size Accuracy (SS): 69.3. Result ID: 69cb6fef47f34ce924b33503.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

### Qwen2.5-VL-7B-Instruct

Shot Size Accuracy (SS): 69.1. Result ID: 69cb6fef47f34ce924b334ed.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

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

### InternVL3-38B

Shot Size Accuracy (SS): 68. Result ID: 69cb6fef47f34ce924b334fd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

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

### InternVL2.5-38B

Shot Size Accuracy (SS): 67.8. Result ID: 69cb6fef47f34ce924b334fc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bab1e209131cea89c7f.md>)

[Implementation](<https://github.com/OpenGVLab/InternVL>)

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

### Qwen2.5-VL-32B-Instruct

Shot Size Accuracy (SS): 62.3. Result ID: 69cb6fef47f34ce924b334fe.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

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

### InternVL3-8B

Shot Size Accuracy (SS): 62.1. Result ID: 69cb6fef47f34ce924b334f3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

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

### InternVL3-14B

Shot Size Accuracy (SS): 59.6. Result ID: 69cb6fef47f34ce924b334f4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

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

### LLaVA-Onevision-Qwen2-7B-Ov-Chat

Shot Size Accuracy (SS): 58.4. Result ID: 69cb6fef47f34ce924b334f0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa756f1e209131cea67e58.md>)

[Implementation](<https://llava-vl.github.io/blog/llava-onevision>)

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

### Gemini-2.5-flash-preview-04-17

Shot Size Accuracy (SS): 57.7. Result ID: 69cb6fef47f34ce924b33502.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

### LLaVA-Video-7B-Qwen2

Shot Size Accuracy (SS): 56.9. Result ID: 69cb6fef47f34ce924b334ef.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

### InternVL2.5-8B

Shot Size Accuracy (SS): 56.3. Result ID: 69cb6fef47f34ce924b334f1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bab1e209131cea89c7f.md>)

[Implementation](<https://github.com/OpenGVLab/InternVL>)

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

### InternVL3-2B

Shot Size Accuracy (SS): 56.3. Result ID: 69cb6fef47f34ce924b334f2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

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

### Qwen2.5-VL-3B-Instruct

Shot Size Accuracy (SS): 54.6. Result ID: 69cb6fef47f34ce924b334ec.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

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

### Internlm-xcomposer2d5-7B

Shot Size Accuracy (SS): 51.1. Result ID: 69cb6fef47f34ce924b334f5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72021e209131cea541e3.md>)

[Implementation](<https://github.com/InternLM/InternLM-XComposer>)

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

### Gemini-2.0-flash

Shot Size Accuracy (SS): 48.9. Result ID: 69cb6fef47f34ce924b33501.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

### VILA1.5-8B

Shot Size Accuracy (SS): 40.6. Result ID: 69cb6fef47f34ce924b334f8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73311e209131cea5b06b.md>)

[Implementation](<https://github.com/NVlabs/VILA>)

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

### VILA1.5-13B

Shot Size Accuracy (SS): 36.7. Result ID: 69cb6fef47f34ce924b334f9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73311e209131cea5b06b.md>)

[Implementation](<https://github.com/NVlabs/VILA>)

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

### LLaVA-NeXT-Video-7B

Shot Size Accuracy (SS): 35.9. Result ID: 69cb6fef47f34ce924b334ee.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

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

### Ovis2-8B

Shot Size Accuracy (SS): 35.9. Result ID: 69cb6fef47f34ce924b334f6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73b21e209131cea5dfaa.md>)

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

### VILA1.5-3B

Shot Size Accuracy (SS): 33.4. Result ID: 69cb6fef47f34ce924b334f7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73311e209131cea5b06b.md>)

[Implementation](<https://github.com/NVlabs/VILA>)

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

### Instructblip-vicuna-7B

Shot Size Accuracy (SS): 27. Result ID: 69cb6fef47f34ce924b334fa.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fb71e209131cea47021.md>)

[Implementation](<https://github.com/salesforce/LAVIS/tree/main/projects/instructblip>)

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

### Instructblip-vicuna-13B

Shot Size Accuracy (SS): 26.8. Result ID: 69cb6fef47f34ce924b334fb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79b31e209131cea7ef72.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fb71e209131cea47021.md>)

[Implementation](<https://github.com/salesforce/LAVIS/tree/main/projects/instructblip>)

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

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