Counting on UHR-Micro (test)
40.98GCMAP-Agent
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MAP-Agentbase_model=Qwen3-VL-8B2026.05 | 40.98 | 43.29 | 29.59 | 17.98 | 32.96 | |
| MAP-Agentbase_model=InternVL3.5-8B2026.05 | 38.42 | 45.94 | 35.46 | 26.35 | 36.54 | |
| DeepEyes2026.05 | 37.88 | 29.24 | 40.03 | 26.84 | 33.5 | |
| Qwen3-VL-235B2026.05 | 33.78 | 4.39 | 23.92 | 18.23 | 20.08 | |
| GPT-5.22026.05 | 33.67 | 14.07 | 23.49 | 25.63 | 24.21 | |
| Qwen2.5-VL-7B2026.05 | 30.59 | 18.22 | 37.34 | 27.96 | 28.53 | |
| Qwen3-VL-30B2026.05 | 29.39 | 5.62 | 31.82 | 25.3 | 23.03 | |
| Gemini-2.5-Pro2026.05 | 28.83 | 33.5 | 32.22 | 20.6 | 28.79 | |
| Qwen3-VL-8B2026.05 | 25.94 | 5.34 | 27.59 | 21.1 | 19.99 | |
| Qwen2-VL-7B2026.05 | 23.83 | 14.08 | 35.05 | 18.38 | 22.84 | |
| GeoChat2026.05 | 22.88 | 14.45 | 51.55 | 24.59 | 28.37 | |
| Claude 3.7 Sonnet2026.05 | 19.88 | 2.81 | 27.61 | 3.23 | 13.38 | |
| InternVL3.5-8B2026.05 | 19.4 | 10.9 | 46.5 | 19.6 | 24.1 | |
| InternVL3.5-38B2026.05 | 3.46 | 7.4 | 12.53 | 13.9 | 9.32 | |
| GeoEyes2026.05 | 3.13 | 0.71 | 0 | 4.84 | 2.17 | |
| GeoLLaVa-8k2026.05 | 0 | 0.19 | 2.16 | 0 | 0.59 |