Visual Reasoning on V* (Pass@1)
88.7Pass@1 ScorePyVision-RL
Evaluation Results
| Method | Links | |
|---|---|---|
| PyVision-RLBase Model=Qwen2.5-VL-7B2026.05 | 88.7 | |
| Mini-o3Base Model=Qwen2.5-VL-7B2026.05 | 88.2 | |
| SFT + AXPOBase Model=Qwen3-VL-8B-Thinking2026.05 | 87.8 | |
| DeepEyesBase Model=Qwen2.5-VL-7B2026.05 | 85.6 | |
| PixelReasonerBase Model=Qwen2.5-VL-7B2026.05 | 84.3 | |
| ThymeBase Model=Qwen2.5-VL-7B2026.05 | 82.2 | |
| DeepEyes-v2Base Model=Qwen2.5-VL-7B2026.05 | 81.8 | |
| Qwen2.5-VL-7B-InstructBase Model=Qwen2.5-VL-7B2026.05 | 78.5 | |
| Qwen3-VL-8B-Thinking (Agent)Base Model=Qwen3-VL-8B-Thinking2026.05 | 77.7 |