ResearchBenchmarksCross-domain demo-to-code on Combination Factor High-ComplexityFollow47.5Success Rate (SR)NESYCR5.916.727.538.3Mar 19, 2026Evaluation ResultsMethodMethodLinksSuccess Rate (SR)Goal Completion (GC)Policy Deviation (PD)NESYCR2026.0347.57517.11Statler2026.0332.573.1261.06Critic-V2026.0327.566.2516.48MoReVQA2026.032562.525.62GPT4V-Robotics2026.031563.1236.46LLM-DM2026.031046.8881.25Demo2Code2026.037.528.7512.5