STEM Multimodal Reasoning on MMMU-Pro
33.74STEM MMMU-Pro ScoreBase
Evaluation results
| Method | Links | |
|---|---|---|
| BaseFine-tuning setting=Base, Backbone=Qwen3-VL-2B2026.07 | 33.74 | |
| GRPOFine-tuning setting=GRPO, Backbone=Qwen3-VL-2B2026.07 | 30.41 | |
| CPOFine-tuning setting=CPO, Backbone=Qwen3-VL-2B2026.07 | 29.48 | |
| GSPOFine-tuning setting=GSPO, Backbone=Qwen3-VL-2B2026.07 | 27.69 | |
| LoRAFine-tuning setting=LoRA, Backbone=Qwen3-VL-2B2026.07 | 27.65 | |
| FFTFine-tuning setting=FFT, Backbone=Qwen3-VL-2B2026.07 | 23.35 |