Share your thoughts, 1 month free Claude Pro on usSee more

Accuracy on RoboSpatial Spatial Reasoning

76.4Accuracy

ViGoRL-7b (Ours)

Updated 25d ago

Evaluation Results

Method	Links
ViGoRL-7b (Ours) 2025.05		76.4
GPT-4o 2025.05		76.2
H-GRPO 2026.06		70.2
Qwen2.5VL-3B + Vanilla GRPO 2025.05		69.7
Molmo 2025.05		67.1
ViGoRL-3b (Ours) 2025.05		67.1
ViGoRL 2026.06		67.1
SpatialRGPT-8B 2026.04		66.7
RoboPIN 2026.06		66.6
OpenVLThinkerV2 2026.04		63.2
G2VLM 2026.04		62.7
Qwen2.5VL-3B + SFT direct 2025.05		62.3
Mimo-Embodied 2026.06		61.8
MiMo-Embodied 2025.11		61.76
OneThinker-8B 2026.04		61.7
Qwen3-VL GRPO 2026.04		61.7
Qwen3-VL GDPO 2026.04		61.7
SpaceLLaVA-13B 2026.04		61
Qwen3-VL-Instruct 2026.04		60.6
Gemini-2.5-Pro 2026.06		59.9
Gemini2.5-Pro 2025.11		59.87
RoboBrain2.0-7B 2026.04		59.6
Qwen3-VL 2026.06		59.4
RoboRefer-8B-SFT 2026.04		58.3
Reason-RFT 2026.06		57.7
Pelican-VL 2026.06		57.5
GPT-5 2026.04		55.5
Qwen2.5VL-3B 2025.05		54.4
Qwen2.5-VL 2026.06		54.4
RoboBrain-2.0 2025.11		54.23
RoboBrain2.0 2026.06		54.2
GPT-5 2026.06		54.1
Claude-Sonnet-4 2025.11		53.72
R1-VL 2026.06		52.5
RoboBrain-1.0 2025.11		51.53
InternVL3.5 2026.06		51.1
MiMo-VL 2025.11		50.05
Qwen2.5-VL 2025.11		49.33
Gemini 2.5 Pro 2026.04		47.5
Embodied-R1 2026.06		47.4
Qwen-VL-Max 2025.11		46.92
Qwen2.5-VL-7B 2025.05		46.7
GPT-4o 2025.11		44.42
VeBrain 2026.06		42.5
VeBrain 2025.11		42.48
Cosmos-Reason1 2025.11		38.81
InternVL3.5 2025.11		37.77
Magma 2025.11		33.71