Spatial Reasoning on RoboSpatial-VQ
79.3AccuracyGR-ER 1.5 Thinking
Evaluation Results
| Method | Links | |
|---|---|---|
| GR-ER 1.5 ThinkingAccess Mode=API call only2026.05 | 79.3 | |
| Qwen3-VL-8BAccess Mode=Open weights only2026.05 | 73.7 | |
| Molmo2-ERAccess Mode=Open weights, Open data, Open code2026.05 | 73.4 | |
| Gemini 2.5 ProAccess Mode=API call only2026.05 | 71.3 | |
| GPT-5-miniAccess Mode=API call only2026.05 | 70.7 | |
| GPT-5Access Mode=API call only2026.05 | 69.3 | |
| LLaVA-OV-7BAccess Mode=Open weights only2026.05 | 68.9 | |
| Qwen3-VL-4BAccess Mode=Open weights only2026.05 | 62.7 | |
| InternVL3.5-8BAccess Mode=Open weights only2026.05 | 62.7 | |
| InternVL3.5-4BAccess Mode=Open weights only2026.05 | 61 | |
| Molmo2Access Mode=Open weights, Open data, Open code2026.05 | 57.9 | |
| GR-ER 1.5Access Mode=API call only2026.05 | 57 |