multimodal Multimodal · llm-stats.com
MMMU
Massive Multi-discipline Multimodal Understanding with college-level image reasoning.
5
Modelli testati
Qwen3.6 Plus
Modello di punta
86%
Punteggio migliore
Alibaba
Provider
| # | Modello | MMMU |
|---|---|---|
| 1 | Qwen3.6 Plus Alibaba | 86% |
| 2 | GPT-5.1 OpenAI | 85.4% |
| 3 | GPT-5 OpenAI | 84.2% |
| 4 | Qwen3.5 122B-A10B Alibaba | 83.9% |
| 5 | o3 OpenAI | 82.9% |
Dati benchmark aggregati da llm-stats.com. Vedi Metodologia.