Single-image spatial reasoning on ERQA 2025
63AccuracyNo-tool
Evaluation Results
| Method | Links | |
|---|---|---|
| No-toolModel=Qwen3.5-397B-A17B2026.06 | 63 | |
| SpatialClawModel=Qwen3.5-397B-A17B2026.06 | 62 | |
| SpatialClawModel=Qwen3.6-27B2026.06 | 61.5 | |
| SpatialClawModel=Gemma4-31B2026.06 | 61.3 | |
| No-toolModel=Qwen3.5-122B-A10B2026.06 | 61 | |
| No-toolModel=Qwen3.6-27B2026.06 | 59.5 | |
| No-toolModel=Gemma4-31B2026.06 | 58.3 | |
| SpatialClawModel=Qwen3.6-35B-A3B2026.06 | 58.2 | |
| SpatialClawModel=Qwen3.5-122B-A10B2026.06 | 57.5 | |
| No-toolModel=Qwen3.6-35B-A3B2026.06 | 57.2 | |
| No-toolModel=Gemma4-26B-A4B2026.06 | 56 | |
| SpatialClawModel=Gemma4-26B-A4B2026.06 | 56 |