Fine-Grained Perception & Understanding on MMBench en (dev)
84.19AccuracyThinkLite-VL-7B
Evaluation results
| Method | Links | |
|---|---|---|
| ThinkLite-VL-7BModel Scale=7B, Training Strategy=Representative Reasoning Model2026.04 | 84.19 | |
| Qwen2.5-VL-7B + SSL-R1Model Scale=7B, Training Strategy=+ SSL-R12026.04 | 84.11 | |
| Qwen2.5-VL-7B + Visual JigsawModel Scale=7B, Training Strategy=+ Visual Jigsaw2026.04 | 83.68 | |
| Qwen2.5-VL-7BModel Scale=7B, Training Strategy=Baseline2026.04 | 83.33 | |
| LLaVA-Critic-R1-7BModel Scale=7B, Training Strategy=Representative Reasoning Model2026.04 | 83.16 | |
| VL-Cogito-7BModel Scale=7B, Training Strategy=Representative Reasoning Model2026.04 | 82.98 | |
| Qwen2.5-VL-3B + SSL-R1Model Scale=3B, Training Strategy=+ SSL-R12026.04 | 78.52 | |
| Qwen2.5-VL-3B + Visual JigsawModel Scale=3B, Training Strategy=+ Visual Jigsaw2026.04 | 78.44 | |
| Qwen2.5-VL-3BModel Scale=3B, Training Strategy=Baseline2026.04 | 77.15 |