Embodied Reasoning and Question Answering on ERQA
65ScoreGemini
Evaluation Results
| Method | Links | |
|---|---|---|
| GeminiModel Version/Scale=3.0 Pro2026.04 | 65 | |
| HY-EmbodiedModel Version/Scale=0.5 MoE A32B2026.04 | 62.3 | |
| SeedModel Version/Scale=2.02026.04 | 61.8 | |
| QwenModel Version/Scale=3.5 A17B2026.04 | 61 | |
| KimiModel Version/Scale=K2.52026.04 | 59.8 | |
| Gemini 2.5 Pro2026.07 | 55.7 | |
| RoboDecision-8BCategory=Embodied Baselines2026.05 | 54.5 | |
| GPT 5.42026.07 | 50.5 | |
| GPT-5Category=General Baselines2026.05 | 49.95 | |
| Gemini-2.5-ProCategory=General Baselines2026.05 | 48.7 | |
| Gemini2.5-proAccess=Closed-source2026.03 | 48.3 | |
| Vero Q3T-8BRL=Qwen3VL, Initial Model=8B Think2026.04 | 47.2 | |
| GPT-4oAccess=Closed-source2026.03 | 47 | |
| MiMo-Embodied-7BAccess=Embodied Brain MLLMs, Parameters=7B2026.03 | 46.8 | |
| Q3VL 8B-ThkRL=N/A, Initial Model=N/A2026.04 | 46.8 | |
| RynnBrainModel Parameters=8B2026.06 | 46.8 | |
| RynnBrain 8B2026.07 | 46.8 | |
| MiMo Embodied-7B2026.07 | 46.8 | |
| ACE-Brain 0.5-8B2026.07 | 46.3 | |
| Cosmos3 Nano2026.07 | 46 | |
| Embodied R1.52026.07 | 46 | |
| Qwen3-VL-8B-Inst.Access=Open-source general-purpose, Parameters=8B2026.03 | 45.8 | |
| Q3VL 8B-InsRL=N/A, Initial Model=N/A2026.04 | 45.8 | |
| GPT-5 NanoRL=N/A, Initial Model=N/A2026.04 | 45.5 | |
| VestaModel Parameters=8B2026.06 | 44.9 | |
| Qwen3-VLModel Parameters=8B2026.06 | 44.8 | |
| RoboBrain2.5-8BAccess=Embodied Brain MLLMs, Parameters=8B2026.03 | 44.3 | |
| RoboBrain 2.5Model Parameters=8B, Evaluation Code=Current Paper's2026.06 | 44 | |
| MiMoVL 7B-RLRL=MiMoVL, Initial Model=7B-SFT2026.04 | 43.5 | |
| Mo2-O 7BRL=SFT Only, Initial Model=N/A2026.04 | 43.5 | |
| Vero Q3I-8BRL=Qwen3VL, Initial Model=8B Inst2026.04 | 43.2 | |
| RoboBrain2.0-7BAccess=Embodied Brain MLLMs, Parameters=7B2026.03 | 42.5 | |
| Vero Q25-7BRL=Qwen25VL, Initial Model=7B Inst2026.04 | 42.5 | |
| Qwen3-VL-8B-InstructCategory=General Baselines2026.05 | 42.5 | |
| Q25VL 7B-InsRL=N/A, Initial Model=N/A2026.04 | 41.8 | |
| InternVL3.5-8BAccess=Open-source general-purpose, Parameters=8B2026.03 | 41.5 | |
| ACE-Brain-0-8BAccess=Embodied Brain MLLMs, Parameters=8B2026.03 | 41.5 | |
| ACE-Brain 0-8B2026.07 | 41.5 | |
| Vlaser-8BAccess=Embodied Brain MLLMs, Parameters=8B2026.03 | 41 | |
| Claude Sonnet 4.62026.07 | 41 | |
| Vlaser 8B2026.07 | 41 | |
| Vero Mi-7BRL=MiMoVL, Initial Model=7B-SFT2026.04 | 40.5 | |
| VeBrain-7BAccess=Embodied Brain MLLMs, Parameters=7B2026.03 | 40.3 | |
| Pelican-VL-7BAccess=Embodied Brain MLLMs, Parameters=7B2026.03 | 39.8 | |
| Pelican VL-7B2026.07 | 39.8 | |
| RoboBrain-7B-2.0Category=Embodied Baselines2026.05 | 39.44 | |
| RoboBrain 2.0-7B2026.07 | 39.3 | |
| Qwen2.5-VL-7B-Inst.Access=Open-source general-purpose, Parameters=7B2026.03 | 38.8 | |
| MiMo-VL-7BAccess=Open-source general-purpose, Parameters=7B2026.03 | 37.8 | |
| InternVL3-8BAccess=Open-source general-purpose, Parameters=8B2026.03 | 35.3 | |
| VeBrain 7B2026.07 | 33.3 | |
| Qwen3-VL-2B-Inst.Access=Open-source general-purpose, Parameters=2B2026.03 | 28.3 | |
| RoboBrain 2.5-8B2026.07 | 22.3 |