Plan-grounded Visual Question Answering on InstructionVidDial (test)
33.65ROUGE-LVIGiA
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| VIGiA2026.02 | 33.65 | 94.02 | |
| InternVL 3.52026.02 | 28.31 | 91.4 | |
| LLaVA-OV2026.02 | 27.02 | 77.38 | |
| LLaVA-1.52026.02 | 26.77 | 64.11 | |
| Qwen 2.5 VL2026.02 | 26.39 | 75.14 | |
| Qwen 3 VL2026.02 | 24.72 | 91.21 | |
| Idefics22026.02 | 21.22 | 44.49 | |
| MM-PlanLLM2026.02 | 7.49 | 0.37 |