Spatial Reasoning on ERQA
55.7AccuracyGemini-2.5-Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-2.5-ProParams=-2026.06 | 55.7 | |
| GPT-5Params=-2026.06 | 54.5 | |
| Mimo-EmbodiedParams=7B2026.06 | 46.8 | |
| RoboPINParams=4B2026.06 | 44 | |
| Qwen3-VL-8BModel Parameters=8B2026.06 | 43 | |
| Ouro-Spatial-8BModel Parameters=8B2026.06 | 42.5 | |
| Qwen3-VLParams=4B2026.06 | 41.8 | |
| Qwen3-VL-4BModel Parameters=4B2026.06 | 41.2 | |
| InternVL3.5Params=8B2026.06 | 41 | |
| Ouro-Spatial-4BModel Parameters=4B2026.06 | 40.5 | |
| Pelican-VLParams=7B2026.06 | 39.8 | |
| AVLM-2BModel Scale=2B2026.06 | 39.75 | |
| RoboBrain2.0Params=7B2026.06 | 38.5 | |
| VeBrainParams=7B2026.06 | 37.3 | |
| Qwen2.5-Omni-3BModel Scale=3B2026.06 | 36.75 | |
| Gemma-4-E2B-itModel Scale=4B2026.06 | 35.5 | |
| Embodied-R1Params=3B2026.06 | 35.2 |