Multimodal Reasoning on PhD
90.57CPBYE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| BYEModel=Qwen3-VL2026.01 | 90.57 | 64.37 | |
| TCAP (Ours)Model=Qwen3-VL2026.01 | 90.27 | 0 | |
| Vanilla FTModel=Qwen3-VL2026.01 | 89.8 | 87.67 | |
| TCAP (Ours)Model=InternVL2.52026.01 | 89.1 | 0 | |
| Vanilla FTModel=InternVL2.52026.01 | 89 | 85.33 | |
| Vanilla FTModel=LLaVA-NeXT2026.01 | 87.77 | 94.77 | |
| Random DropModel=Qwen3-VL2026.01 | 87.5 | 90.03 | |
| TCAP (Ours)Model=LLaVA-NeXT2026.01 | 87.23 | 0.67 | |
| Random DropModel=InternVL2.52026.01 | 86 | 82.13 | |
| Random DropModel=LLaVA-NeXT2026.01 | 84.87 | 93.03 | |
| SampDetoxModel=InternVL2.52026.01 | 82.97 | 74.87 | |
| BYEModel=InternVL2.52026.01 | 82.37 | 14.4 | |
| SampDetoxModel=Qwen3-VL2026.01 | 81.7 | 71.73 | |
| SampDetoxModel=LLaVA-NeXT2026.01 | 76.63 | 84.4 | |
| BYEModel=LLaVA-NeXT2026.01 | 0 | 100 |