Massive Multi-discipline Multimodal Understanding on MMMU (val)
65.34MMMU ScoreETTC
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ETTCAggregation Method=ETTC2026.05 | 65.34 | — | — | |
| Qwen-72BModel Scale=72B2026.05 | 64.18 | — | — | |
| Qwen-32BModel Scale=32B2026.05 | 59.04 | — | — | |
| VotingAggregation Method=Voting2026.05 | 58.63 | — | — | |
| Qwen3-VL-8BSetup=Zero-shot, Base Model=Qwen3-VL-8B2026.03 | 53.4 | 100 | — | |
| AverageAggregation Method=Mean2026.05 | 52.79 | — | — | |
| ResAdaptSetup=Zero-shot, Base Model=Qwen2.5-VL-7B2026.03 | 51 | 29 | — | |
| Qwen2.5-VL-7BSetup=Zero-shot, Base Model=Qwen2.5-VL-7B2026.03 | 50.9 | 100 | — | |
| ResAdapt-RLSetup=Zero-shot, Base Model=Qwen2.5-VL-7B, RL fine-tuning=true2026.03 | 50.9 | 29 | — | |
| ResAdaptSetup=Zero-shot, Base Model=Qwen3-VL-8B2026.03 | 50.9 | 29 | — | |
| ToMeSetup=Zero-shot, Base Model=Qwen3-VL-8B2026.03 | 50.6 | 50 | — | |
| Qwen-7BModel Scale=7B2026.05 | 50.53 | — | — | |
| VisionZipSetup=Zero-shot, Base Model=Qwen3-VL-8B2026.03 | 50.3 | 50 | — | |
| ToMeSetup=Zero-shot, Base Model=Qwen2.5-VL-7B2026.03 | 49.6 | 50 | — | |
| Random DropSetup=Zero-shot, Base Model=Qwen2.5-VL-7B2026.03 | 49 | 50 | — | |
| Random DropSetup=Zero-shot, Base Model=Qwen3-VL-8B2026.03 | 48.7 | 50 | — | |
| VisionZipSetup=Zero-shot, Base Model=Qwen2.5-VL-7B2026.03 | 48.6 | 50 | — | |
| Qwen-3BModel Scale=3B2026.05 | 37.41 | — | — | |
| InternVL3.5-1B (+GNDPO)Model Scale=1B, Optimization Method=GNDPO2026.06 | — | — | 47.6 | |
| InternVL3.5-1B (+GSPO)Model Scale=1B, Optimization Method=GSPO2026.06 | — | — | 39.9 | |
| InternVL3.5-1B (+OPD)Model Scale=1B, Optimization Method=OPD2026.06 | — | — | 47.1 | |
| InternVL3.5-1B (Base (Instruct))Model Scale=1B, Optimization Method=Base (Instruct)2026.06 | — | — | 48.6 | |
| InternVL3.5-2B (+GNDPO)Model Scale=2B, Optimization Method=GNDPO2026.06 | — | — | 57.7 | |
| InternVL3.5-2B (+GSPO)Model Scale=2B, Optimization Method=GSPO2026.06 | — | — | 54.7 | |
| InternVL3.5-2B (+OPD)Model Scale=2B, Optimization Method=OPD2026.06 | — | — | 57.4 | |
| InternVL3.5-2B (Base (Instruct))Model Scale=2B, Optimization Method=Base (Instruct)2026.06 | — | — | 53 | |
| InternVL3.5-4B (+GNDPO)Model Scale=4B, Optimization Method=GNDPO2026.06 | — | — | 68.3 | |
| InternVL3.5-4B (+GSPO)Model Scale=4B, Optimization Method=GSPO2026.06 | — | — | 67.2 | |
| InternVL3.5-4B (+OPD)Model Scale=4B, Optimization Method=OPD2026.06 | — | — | 67.2 | |
| InternVL3.5-4B (Base (Instruct))Model Scale=4B, Optimization Method=Base (Instruct)2026.06 | — | — | 64.3 | |
| InternVL3.5-8BModel Scale=8B, Optimization Method=N/A2026.06 | — | — | 73.4 |