STEM Multimodal Reasoning on MathVision (test)
19.93ScoreCPO
Evaluation results
| Method | Links | |
|---|---|---|
| CPOFine-tuning setting=CPO, Backbone=Qwen3-VL-2B2026.07 | 19.93 | |
| GSPOFine-tuning setting=GSPO, Backbone=Qwen3-VL-2B2026.07 | 18.19 | |
| GRPOFine-tuning setting=GRPO, Backbone=Qwen3-VL-2B2026.07 | 17.43 | |
| LoRAFine-tuning setting=LoRA, Backbone=Qwen3-VL-2B2026.07 | 16.91 | |
| FFTFine-tuning setting=FFT, Backbone=Qwen3-VL-2B2026.07 | 15.99 | |
| BaseFine-tuning setting=Base, Backbone=Qwen3-VL-2B2026.07 | 14.87 |