Embodied Visual Question Answering on ERQA
59AccuracyGPT-5
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5Access Mode=API call only2026.05 | 59 | |
| GPT-5-miniAccess Mode=API call only2026.05 | 57.3 | |
| Gemini 2.5 ProAccess Mode=API call only2026.05 | 56 | |
| Gemini-2.5-ProModel Category=General-Purpose VLMs2026.06 | 55.7 | |
| GR-ER 1.5 ThinkingAccess Mode=API call only2026.05 | 54.8 | |
| Qwen3-VL-32B-Instruct + SpatialBoostEncoder Enhancement=SpatialBoost2026.03 | 51.5 | |
| GPT-5.4Model Category=General-Purpose VLMs2026.06 | 50.5 | |
| InternVL3-38B + SpatialBoostEncoder Enhancement=SpatialBoost2026.03 | 49.3 | |
| Qwen3-VL-32B-Instruct2026.03 | 48.8 | |
| Gemini-Robotics-ER-1.5Model Category=Embodied VLMs2026.06 | 48.5 | |
| GR-ER 1.5Access Mode=API call only2026.05 | 47 | |
| Molmo2-ERAccess Mode=Open weights, Open data, Open code2026.05 | 46.8 | |
| Mimo-EmbodiedModel Category=Embodied VLMs2026.06 | 46.8 | |
| Molmo2Access Mode=Open weights, Open data, Open code2026.05 | 46.3 | |
| Embodied-R1.5Model Category=Ours2026.06 | 46 | |
| Qwen3-VL-8BParameters=8B2026.06 | 43.5 | |
| Ours-8BParameters=8B2026.06 | 43.5 | |
| InternVL3-38B2026.03 | 42.8 | |
| Qwen3-VL-8BModel Category=General-Purpose VLMs2026.06 | 41.8 | |
| Qwen3-VL-8BAccess Mode=Open weights only2026.05 | 41.5 | |
| InternVL3.5-8BAccess Mode=Open weights only2026.05 | 41 | |
| InternVL3.5-8BModel Category=General-Purpose VLMs2026.06 | 41 | |
| Molmo-D-7BModel Category=General-Purpose VLMs2026.06 | 40.6 | |
| InternVLA-M1Parameters=M12026.06 | 40.35 | |
| GR3D-8BTraining Stage=Stage 12026.05 | 40.25 | |
| LLaVA-OV-7BAccess Mode=Open weights only2026.05 | 39.8 | |
| Pelican-VL-1.0Model Category=Embodied VLMs2026.06 | 39.8 | |
| Qwen3-VL-4BAccess Mode=Open weights only2026.05 | 39.5 | |
| Embodied-R1-7BParameters=7B2026.06 | 39.1 | |
| GR3D-8BTraining Stage=Stage 22026.05 | 38.5 | |
| RoboBrain-2.0-7BModel Category=Embodied VLMs2026.06 | 38.5 | |
| VeBrainModel Category=Embodied VLMs2026.06 | 37.3 | |
| InternVL3.5-4BAccess Mode=Open weights only2026.05 | 37 | |
| NVILA-Lite-8B2026.05 | 36.25 | |
| Embodied-R1Model Category=Embodied VLMs2026.06 | 35.2 | |
| RoboBrain2.0-7BParameters=7B2026.06 | 34.25 | |
| GPT-4oModel Category=General-Purpose VLMs2026.06 | 32.5 | |
| InternVL3-8BParameters=8B2026.06 | 32.08 | |
| MagmaModel Category=Embodied VLMs2026.06 | 25.4 |