Referring Expression Comprehension on HumanRef
87.1PrecisionVLM-FO1 3B
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| VLM-FO1 3BFPS=null2025.12 | 87.1 | 83.3 | 82.6 | — | — | — | — | — | |
| WeDetect-Ref 4BFPS=5.32025.12 | 86.3 | 87.1 | 81.8 | 64.1 | — | — | — | — | |
| RexSeek 7BFPS=null2025.12 | 85.8 | 85.9 | 82.4 | 54.1 | — | — | — | — | |
| WeDetect-Ref 2BFPS=6.62025.12 | 84.7 | 85.1 | 79.8 | 61 | — | — | — | — | |
| Rex-Omni 3BFPS=null2025.12 | 79.3 | 80.1 | 75.6 | — | — | — | — | — | |
| Qwen3-VL 4BFPS=0.42025.12 | 76.7 | 65.9 | 67.8 | 39.1 | — | — | — | — | |
| ChatRex 7BFPS=null2025.12 | 72.2 | 50.4 | 55.6 | 0 | — | — | — | — | |
| Qwen3-VL 2BFPS=0.62025.12 | 69.7 | 58.2 | 60.2 | 20.6 | — | — | — | — | |
| Qwen2.5-VL 7BFPS=null2025.12 | 68.5 | 52.5 | 56.2 | 7.1 | — | — | — | — | |
| InternVL2.5-8BFPS=null2025.12 | 37.9 | 29.8 | 31.9 | 54.9 | — | — | — | — | |
| Grounding-DINO-LFPS=3.12025.12 | 33.1 | 75.2 | 23.3 | — | — | — | — | — | |
| BagelOutput Format=bbox2026.07 | — | — | — | — | — | — | — | 74.6 | |
| DeepSeek-VL2-Small2026.05 | — | — | — | — | 72 | 46.5 | 64.7 | — | |
| DeepSeek-VL2-Tiny2026.05 | — | — | — | — | 39.1 | 16.9 | 31.4 | — | |
| Grounding DINO-Swin-TScore Thresh.=0.252026.05 | — | — | — | — | 28 | 16.5 | 25.2 | — | |
| Grounding DINO-Swin-TOutput Format=bbox2026.07 | — | — | — | — | — | — | — | 25.2 | |
| LocateAnythingOutput Format=bbox2026.07 | — | — | — | — | — | — | — | 78.7 | |
| LocateAnything-3B2026.05 | — | — | — | — | 82.9 | 68.8 | 78.7 | — | |
| MiMo-VL-7B2026.05 | — | — | — | — | 77.6 | 26.4 | 63.4 | — | |
| OVIS2.5-2B2026.05 | — | — | — | — | 70.6 | 12.3 | 50 | — | |
| Qwen3-VL-4B2026.05 | — | — | — | — | 77.7 | 54.9 | 71.1 | — | |
| Qwen3-VL-8B2026.05 | — | — | — | — | 78.6 | 55.7 | 72 | — | |
| Qwen3-VL-8B-InstructOutput Format=bbox2026.07 | — | — | — | — | — | — | — | 70.4 | |
| Qwen3.5-9BOutput Format=bbox2026.07 | — | — | — | — | — | — | — | 71.7 | |
| Rex-OmniOutput Format=bbox2026.07 | — | — | — | — | — | — | — | 79.9 | |
| Rex-Omni-3B2026.05 | — | — | — | — | 85.4 | 65.4 | 79.9 | — | |
| SEED1.5-VL2026.05 | — | — | — | — | 88.2 | 60 | 81.6 | — | |
| SenseNova-VisionOutput Format=bbox2026.07 | — | — | — | — | — | — | — | 80.2 |