Emotion prediction on WebEmo-2 ID VSA, VECBench
75.9AccuracyQwen3-VL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3-VLModel Category=Open-Source MLLM2025.12 | 75.9 | 75.71 | |
| EmoCaliberModel Category=Emotion-Oriented MLLM2025.12 | 75.8 | 75.78 | |
| EmoVITModel Category=Emotion-Oriented MLLM2025.12 | 74.4 | 74.67 | |
| GPT-5Model Category=Proprietary MLLM2025.12 | 73.7 | 73.51 | |
| InternVL3.5Model Category=Open-Source MLLM2025.12 | 70.3 | 69.47 | |
| Qwen2.5-VLModel Category=Open-Source MLLM2025.12 | 69.5 | 69.64 | |
| Emotion-LLaMAModel Category=Emotion-Oriented MLLM2025.12 | 44.9 | 48.21 |