ResearchBenchmarksLong-horizon tasks on Minecraft GoldFollow21.69Success Rate (SR)EvoAgent-0.86764.988710.84516.7013Feb 9, 2025Evaluation ResultsMethodMethodLinksSuccess Rate (SR)Efficiency Error (EE)EvoAgent2025.0221.6930.48Optimus-12025.0210.628.03Jarvis-12025.028.849.76LS-Imagine2025.026.6110.69DreamerV32025.026.578.05PPO2025.0200GPT-4V2025.0200