Continuous Subjective Prediction on A3-Dataset
0.88Aesthetic SRCCA3-Align
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| A3-AlignModel Category=Ours, Thinking Mode=false2026.03 | 0.88 | 0.88 | 0.838 | 0.836 | |
| Claude Opus 4.1Model Category=Closed-source Models, Thinking Mode=false2026.03 | 0.772 | 0.749 | 0.749 | 0.751 | |
| Gemini 2.5 pro-thinkingModel Category=Closed-source Models - Thinking, Thinking Mode=true2026.03 | 0.759 | 0.762 | 0.764 | 0.743 | |
| Gemini 2.5 proModel Category=Closed-source Models, Thinking Mode=false2026.03 | 0.743 | 0.744 | 0.704 | 0.686 | |
| ChatGPT 5 highModel Category=Closed-source Models - Thinking, Thinking Mode=true2026.03 | 0.74 | 0.703 | 0.764 | 0.714 | |
| ChatGPT o3Model Category=Closed-source Models - Thinking, Thinking Mode=true2026.03 | 0.736 | 0.699 | 0.763 | 0.715 | |
| ChatGPT 5Model Category=Closed-source Models, Thinking Mode=false2026.03 | 0.732 | 0.698 | 0.752 | 0.703 | |
| Claude Sonnet 4.5-thinkingModel Category=Closed-source Models - Thinking, Thinking Mode=true2026.03 | 0.732 | 0.697 | 0.713 | 0.696 | |
| Claude Sonnet 4.5Model Category=Closed-source Models, Thinking Mode=false2026.03 | 0.725 | 0.691 | 0.689 | 0.677 | |
| Gemini 2.5 flash-thinkingModel Category=Closed-source Models - Thinking, Thinking Mode=true2026.03 | 0.725 | 0.702 | 0.624 | 0.629 | |
| Doubao 1.5 vision pro-thinkingModel Category=Closed-source Models - Thinking, Thinking Mode=true2026.03 | 0.724 | 0.673 | 0.716 | 0.66 | |
| Claude Opus 4.1-thinkingModel Category=Closed-source Models - Thinking, Thinking Mode=true2026.03 | 0.722 | 0.701 | 0.754 | 0.736 | |
| ChatGPT o4 mini highModel Category=Closed-source Models - Thinking, Thinking Mode=true2026.03 | 0.711 | 0.667 | 0.72 | 0.702 | |
| Gemini 2.5 flashModel Category=Closed-source Models, Thinking Mode=false2026.03 | 0.704 | 0.676 | 0.626 | 0.622 | |
| Doubao 1.5 vision proModel Category=Closed-source Models, Thinking Mode=false2026.03 | 0.699 | 0.663 | 0.587 | 0.539 | |
| Grok 4Model Category=Closed-source Models, Thinking Mode=false2026.03 | 0.698 | 0.674 | 0.69 | 0.672 | |
| Doubao seed 1.6-thinkingModel Category=Closed-source Models - Thinking, Thinking Mode=true2026.03 | 0.688 | 0.651 | 0.729 | 0.704 | |
| Mistral medium 3.1Model Category=Closed-source Models, Thinking Mode=false2026.03 | 0.686 | 0.66 | 0.604 | 0.587 | |
| ChatGPT 4oModel Category=Closed-source Models, Thinking Mode=false2026.03 | 0.68 | 0.648 | 0.781 | 0.756 | |
| Doubao seed 1.6 visionModel Category=Closed-source Models, Thinking Mode=false2026.03 | 0.68 | 0.633 | 0.527 | 0.458 | |
| Gemma-3-27B-itModel Category=Open-source Models, Thinking Mode=false2026.03 | 0.677 | 0.657 | 0.66 | 0.66 | |
| ChatGPT 4.1Model Category=Closed-source Models, Thinking Mode=false2026.03 | 0.676 | 0.64 | 0.703 | 0.645 | |
| Qwen3-VL-235B-A22B-thinkingModel Category=Open-source Models - Thinking, Thinking Mode=true2026.03 | 0.667 | 0.613 | 0.669 | 0.651 | |
| Glm-4.5v-106BModel Category=Open-source Models, Thinking Mode=false2026.03 | 0.652 | 0.618 | 0.661 | 0.65 | |
| Qwen3-VL-32B-InstructModel Category=Open-source Models, Thinking Mode=false2026.03 | 0.64 | 0.617 | 0.728 | 0.701 | |
| Llama-4-Scout-109BModel Category=Open-source Models, Thinking Mode=false2026.03 | 0.636 | 0.613 | 0.565 | 0.562 | |
| Qwen3-VL-32B-thinkingModel Category=Open-source Models - Thinking, Thinking Mode=true2026.03 | 0.607 | 0.599 | 0.642 | 0.618 | |
| Claude Haiku 4.5Model Category=Closed-source Models, Thinking Mode=false2026.03 | 0.571 | 0.542 | 0.68 | 0.658 | |
| Claude Haiku 4.5-thinkingModel Category=Closed-source Models - Thinking, Thinking Mode=true2026.03 | 0.568 | 0.555 | 0.729 | 0.705 | |
| Qwen3-VL-8B-InstructModel Category=Open-source Models, Thinking Mode=false2026.03 | 0.564 | 0.562 | 0.533 | 0.528 | |
| Qwen2.5-VL-72B-InstructModel Category=Open-source Models, Thinking Mode=false2026.03 | 0.544 | 0.512 | 0.628 | 0.597 | |
| Llama-4-Maverick-400BModel Category=Open-source Models, Thinking Mode=false2026.03 | 0.49 | 0.478 | 0.539 | 0.519 |