Audiovisual Understanding & Reasoning on World-Sense
65.1ScoreGemini 2.5 Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini 2.5 ProArchitecture=Native MLLM2026.02 | 65.1 | |
| TAC-VArchitecture=Describer + Text-only Reasoner, Reasoner=Gemini32026.02 | 58.6 | |
| Gemini2.5-Pro-MTSSModel Category=Proprietary, Scripting Paradigm=MTSS2026.04 | 49.81 | |
| OmniVinciArchitecture=Native MLLM2026.02 | 48.2 | |
| TAC-VArchitecture=Describer + Text-only Reasoner, Reasoner=Qwen32026.02 | 45.7 | |
| Gemini2.5-ProModel Category=Proprietary, Scripting Paradigm=Monolithic2026.04 | 43.32 | |
| Qwen3-Omni-MTSS-FTModel Category=Open-Source, Scripting Paradigm=MTSS, Evaluation Protocol=Fine-tuned2026.04 | 38.75 | |
| VLMArchitecture=Describer + Text-only Reasoner, Reasoner=Qwen32026.02 | 37.4 | |
| AVoCaDOModel Category=Open-Source2026.04 | 35.31 | |
| Qwen3.5-Omni-Flash-MTSSModel Category=Proprietary, Scripting Paradigm=MTSS2026.04 | 33.04 | |
| ASID-Captioner-7BModel Category=Open-Source2026.04 | 32.03 | |
| Qwen3-Omni-MTSSModel Category=Open-Source, Scripting Paradigm=MTSS2026.04 | 31.06 | |
| Qwen3.5-Omni-FlashModel Category=Proprietary, Scripting Paradigm=Monolithic2026.04 | 24.76 | |
| Qwen3-OmniModel Category=Open-Source, Scripting Paradigm=Monolithic2026.04 | 15.69 |