Audiovisual Understanding & Reasoning on Video-Holmes (Score)
59.2ScoreTAC-V
Evaluation Results
| Method | Links | |
|---|---|---|
| TAC-VArchitecture=Describer + Text-only Reasoner, Reasoner=Gemini32026.02 | 59.2 | |
| Qwen3-OmniArchitecture=Native MLLM2026.02 | 57.3 | |
| TAC-VArchitecture=Describer + Text-only Reasoner, Reasoner=Qwen32026.02 | 47.7 | |
| VLMArchitecture=Describer + Text-only Reasoner, Reasoner=Qwen32026.02 | 45.6 |