Spatial Reasoning on SEED-Bench-2-Plus
73AccuracyQwen2.5-VL-72B-Instruct
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-VL-72B-InstructModel=Qwen2.5-VL-72B-Instruct, Training Strategy=Reference2025.10 | 73 | |
| Qwen2.5-VL-32B-InstructModel=Qwen2.5-VL-32B-Instruct, Training Strategy=Reference2025.10 | 72.1 | |
| REWARDMAPModel=Qwen2.5-VL-7B-Instruct, Training Strategy=REWARDMAP2025.10 | 61.96 | |
| Qwen2.5-VL-7B-Instruct + SFT -> RLModel=Qwen2.5-VL-7B-Instruct, Training Strategy=SFT -> RL Baseline2025.10 | 61.59 | |
| Qwen2.5-VL-7B-InstructModel=Qwen2.5-VL-7B-Instruct, Training Strategy=Base Model2025.10 | 60.97 | |
| Qwen2.5-VL-3B-InstructModel=Qwen2.5-VL-3B-Instruct, Training Strategy=Reference2025.10 | 58.86 | |
| Kimi-VL-A3B-InstructModel=Kimi-VL-A3B-Instruct, Training Strategy=Reference2025.10 | 58.49 |