Multi-modal Understanding on MMMU (test)
62.9Final PerformanceFull RT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Full RTBackbone=Qwen3-VL-8B+2B-R2026.04 | 62.9 | — | — | — | |
| MAIBackbone=Qwen3-VL-2B/8B2026.04 | 62 | 498 | 142 | 260 | |
| 8B(S) BaselineBackbone=Qwen3-VL-8B, Mode=Standard2026.04 | 55.6 | — | — | — | |
| 2B(R) BaselineBackbone=Qwen3-VL-2B, Mode=Reasoning2026.04 | 47.9 | — | — | — | |
| 2B(S) BaselineBackbone=Qwen3-VL-2B, Mode=Standard2026.04 | 44.2 | — | — | — |