Visual Question Answering on VisualProbe Hard
47.2AccuracyTTSP
Evaluation results
| Method | Links | |
|---|---|---|
| TTSPModel=Qwen3-VL-8B-Instruct2026.04 | 47.2 | |
| DeepconfModel=Qwen3-VL-8B-Instruct2026.04 | 43.4 | |
| SCModel=Qwen3-VL-8B-Instruct2026.04 | 42.5 | |
| RTWIModel=Qwen3-VL-8B-Instruct2026.04 | 42.5 | |
| TRTModel=Qwen3-VL-8B-Instruct2026.04 | 40.6 | |
| CISCModel=Qwen3-VL-8B-Instruct2026.04 | 39.6 | |
| Self-Ref.Model=Qwen3-VL-8B-Instruct2026.04 | 38.7 | |
| Self-Cer.Model=Qwen3-VL-8B-Instruct2026.04 | 37.7 | |
| BaseModel=Qwen3-VL-8B-Instruct2026.04 | 34.9 |