Multi-modal Question Answering on MMMU Pros
75AccuracyQwen3.5-27B
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen3.5-27BMode=REASONING, Architecture=Dense, # Total Params=27B, # Activated Params=27B2026.04 | 75 | |
| Qwen3-VL-235B-A22BMode=Thinking, Architecture=MoE, # Total Params=236B, # Activated Params=23B2026.04 | 69.3 | |
| EXAONE 4.5 33BMode=REASONING, Architecture=Dense, # Total Params=33B, # Activated Params=33B2026.04 | 68.6 | |
| Qwen3-VL-32BMode=Thinking, Architecture=Dense, # Total Params=33B, # Activated Params=33B2026.04 | 68.1 | |
| GPT-5 miniMode=REASONING: HIGH, Architecture=-, # Total Params=-, # Activated Params=-2026.04 | 67.3 | |
| MergeMixbaseline=SFT Vision2025.10 | 37.46 | |
| VisionThink-7B2025.10 | 37.27 | |
| SFT Vision2025.10 | 36.7 | |
| Qwen2.5-VL-Ins-7B2025.10 | 36.42 |