# Aesthetic Quality Assessment on UltraVQA

Source: [Aesthetic Quality Assessment on UltraVQA](<https://www.sota2.com/research/sota/aesthetic-quality-assessment-on-ultravqa>)

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

Data updated: 2026-02-26T05:34:35.795Z

Evaluation of Aesthetic Quality Assessment performance on the UltraVQA dataset, using metrics such as Acc@0.5, MAE, SRCC, and PLCC.

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 | Aesthetic Quality Assessment |
| Dataset (including any stated split) | UltraVQA |
| Metric | Accuracy @0.5 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 14 of 14 |
| 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

- [Accuracy @0.5](<https://www.sota2.com/research/sota/aesthetic-quality-assessment-on-ultravqa.md?metric=Accuracy%20%400.5>): Higher is better; unit: PERCENT.

- [MAE](<https://www.sota2.com/research/sota/aesthetic-quality-assessment-on-ultravqa.md?metric=MAE>): Lower is better; unit: SCALAR.

- [SRCC](<https://www.sota2.com/research/sota/aesthetic-quality-assessment-on-ultravqa.md?metric=SRCC>): Higher is better; unit: SCALAR.

- [PLCC](<https://www.sota2.com/research/sota/aesthetic-quality-assessment-on-ultravqa.md?metric=PLCC>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Accuracy @0.5 | Result date | Report paper |
| --- | --- | --- | --- |
| ASO | 85 | Not recorded | 6997eaac09c33900fcd766cf |
| GRPO | 84.7 | Not recorded | 6997eaac09c33900fcd766cf |
| SFT | 74 | Not recorded | 6997eaac09c33900fcd766cf |
| FineVQ | 72.5 | Not recorded | 6997eaac09c33900fcd766cf |
| VideoScoreV2 | 63.7 | Not recorded | 6997eaac09c33900fcd766cf |
| GPT-4.1 | 56.9 | Not recorded | 6997eaac09c33900fcd766cf |
| Gemini-2.5Pro | 56 | Not recorded | 6997eaac09c33900fcd766cf |
| VideoLLaMA3 | 50.2 | Not recorded | 6997eaac09c33900fcd766cf |
| InternVideo-2.5 | 45.7 | Not recorded | 6997eaac09c33900fcd766cf |
| MiniCPMV-4.5 | 44.3 | Not recorded | 6997eaac09c33900fcd766cf |
| Q-Align | 44.3 | Not recorded | 6997eaac09c33900fcd766cf |
| VideoChat-Flash | 43.5 | Not recorded | 6997eaac09c33900fcd766cf |
| Qwen2.5-VL | 41 | Not recorded | 6997eaac09c33900fcd766cf |
| InternVL-3.5 | 40.3 | Not recorded | 6997eaac09c33900fcd766cf |

### ASO

Accuracy @0.5: 85. Result ID: 699b007a5698665ec4925eb9.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

Recorded evaluation configuration: {"Model Category":"Ours","Protocol":"Analytic Score Optimization"}. This record does not assert independent reproduction.

### GRPO

Accuracy @0.5: 84.7. Result ID: 699b007a5698665ec4925eb8.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73101e209131cea5a395.md>)

[Implementation](<https://github.com/deepseek-ai/DeepSeek-Math>)

Recorded evaluation configuration: {"Model Category":"Ours","Protocol":"Group Relative Policy Optimization"}. This record does not assert independent reproduction.

### SFT

Accuracy @0.5: 74. Result ID: 699b007a5698665ec4925eb7.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

Recorded evaluation configuration: {"Model Category":"Ours","Protocol":"Supervised Fine-Tuning"}. This record does not assert independent reproduction.

### FineVQ

Accuracy @0.5: 72.5. Result ID: 699b007a5698665ec4925eb4.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78461e209131cea778fa.md>)

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

### VideoScoreV2

Accuracy @0.5: 63.7. Result ID: 699b007a5698665ec4925eb6.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a07.md>)

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

### GPT-4.1

Accuracy @0.5: 56.9. Result ID: 699b007a5698665ec4925eac.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

Recorded evaluation configuration: {"Model Category":"VLM-APIs"}. This record does not assert independent reproduction.

### Gemini-2.5Pro

Accuracy @0.5: 56. Result ID: 699b007a5698665ec4925ead.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

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

Recorded evaluation configuration: {"Model Category":"VLM-APIs"}. This record does not assert independent reproduction.

### VideoLLaMA3

Accuracy @0.5: 50.2. Result ID: 699b007a5698665ec4925eb0.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79111e209131cea7bf26.md>)

Recorded evaluation configuration: {"Model Category":"Open-source VLMs"}. This record does not assert independent reproduction.

### InternVideo-2.5

Accuracy @0.5: 45.7. Result ID: 699b007a5698665ec4925eb3.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79fd1e209131cea807f5.md>)

[Implementation](<https://github.com/OpenGVLab/InternVideo/tree/main/InternVideo2.5>)

Recorded evaluation configuration: {"Model Category":"Open-source VLMs"}. This record does not assert independent reproduction.

### MiniCPMV-4.5

Accuracy @0.5: 44.3. Result ID: 699b007a5698665ec4925eaf.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b8d1e209131cea8913f.md>)

Recorded evaluation configuration: {"Model Category":"Open-source VLMs"}. This record does not assert independent reproduction.

### Q-Align

Accuracy @0.5: 44.3. Result ID: 699b007a5698665ec4925eb5.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71a31e209131cea523a2.md>)

[Implementation](<https://github.com/Q-Future/Q-Align>)

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

### VideoChat-Flash

Accuracy @0.5: 43.5. Result ID: 699b007a5698665ec4925eb1.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79fd1e209131cea807ed.md>)

Recorded evaluation configuration: {"Model Category":"Open-source VLMs"}. This record does not assert independent reproduction.

### Qwen2.5-VL

Accuracy @0.5: 41. Result ID: 699b007a5698665ec4925eae.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

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

Recorded evaluation configuration: {"Model Category":"Open-source VLMs"}. This record does not assert independent reproduction.

### InternVL-3.5

Accuracy @0.5: 40.3. Result ID: 699b007a5698665ec4925eb2.

[Reporting paper](<https://www.sota2.com/research/paper/6997eaac09c33900fcd766cf.md>)

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

Recorded evaluation configuration: {"Model Category":"Open-source VLMs"}. This record does not assert independent reproduction.

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