Mathematical Reasoning on Overall Combined Benchmarks
58.4Avg@3 ScoreADORA
Evaluation Results
| Method | Links | |
|---|---|---|
| ADORABackbone=Qwen2.5-VL-7B, RL Algorithm=GRPO + ADORA2026.02 | 58.4 | |
| Vision-R1-7B2026.02 | 57.2 | |
| GRPOBackbone=Qwen2.5-VL-7B, RL Algorithm=GRPO2026.02 | 54 | |
| Qwen2.5-VL-7BRL Algorithm=None2026.02 | 51 | |
| ADORABackbone=Internvl3-2b, RL Algorithm=GRPO + ADORA2026.02 | 39.25 | |
| GRPOBackbone=Internvl3-2b, RL Algorithm=GRPO2026.02 | 35.3 | |
| Internvl3-2bRL Algorithm=None2026.02 | 32.35 | |
| ADORABackbone=Gemma3-4b-it, RL Algorithm=GRPO + ADORA2026.02 | 25.28 | |
| GRPOBackbone=Gemma3-4b-it, RL Algorithm=GRPO2026.02 | 24.18 | |
| Gemma3-4b-itRL Algorithm=None2026.02 | 23.88 |