STEM Reasoning on MMMU-Pro Vision
56.3AccuracyQwen-8B-DeltaThinker
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen-8B-DeltaThinker2026.05 | 56.3 | |
| OPD w/ DELTAPROMPTS (10k)Training Data=DELTAPROMPTS (10k)2026.05 | 53.75 | |
| OPD w/ Seed Data (10k)Training Data=Seed Data (10k)2026.05 | 53.64 | |
| Qwen3-VL-8B-Thinking2026.05 | 53.41 |