Video Question Answering on Video QA Benchmarks (test)
74.3MMVUQwen3-VL-235B-Think
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Qwen3-VL-235B-ThinkModel Size=235B2026.05 | 74.3 | 78.98 | 75.78 | 50.87 | 70.18 | |
| Qwen3-VL-235B-InstrModel Size=235B2026.05 | 70.3 | 80.59 | 72.67 | 54.55 | 69.58 | |
| VCapIteration=e2, Base Model=Qwen3-VL-8B2026.05 | 63.1 | 73.92 | 60.44 | 46.35 | 63.82 | |
| VCapIteration=e1, Base Model=Qwen3-VL-8B2026.05 | 62.5 | 74.43 | 58.44 | 46.16 | 63 | |
| Qwen3-VL-8B-InstrModel Size=8B2026.05 | 60.8 | 74.2 | 57 | 45.64 | 60.99 |