List your company

Robotic Failure Analysis on RoboFAC 1.0 (mixed simulated and real-world)

82.74Task Success Rate (Short Horizon)RoboFAC-7B
11.520830.010448.566.9896May 18, 2025

Evaluation results

MethodLinks
RoboFAC-7BParameters=7B2025.0582.7484.9281.7883.2868.9479.1
RoboFAC-3BParameters=3B2025.0581.6684.6779.3283.0263.2976.8
Gemini-2.02025.0563.3253.2345.6748.9141.7251.11
GPT-4o2025.0561.553.8142.4645.8265.8957.42
Qwen-2.5-VL-3BParameters=3B2025.0540.9927.8225.1828.9417.3627.82
Qwen-2.5-VL-7BParameters=7B2025.0514.2611.7338.841850.9627.47