ResearchBenchmarksTask Planning and Program Generation on IMR-Bench Simple Multi-RobotFollow100OCIMR-LLM65.6874.5983.592.41Mar 3, 2026Evaluation ResultsMethodMethodLinksOCSEExeGCRSRIMR-LLMBackbone=GPT-4oBackbone=GPT-4o2026.03100100879487IMR-LLMBackbone=Qwen3-32BBackbone=Qwen3-32B2026.0310093879087LiP-OBackbone=GPT-4oBackbone=GPT-4o2026.039380737473LaMMA-SBackbone=GPT-4oBackbone=GPT-4o2026.037167404533LaMMA-OBackbone=GPT-4oBackbone=GPT-4o2026.037167535546SMART-LLMBackbone=GPT-4oBackbone=GPT-4o2026.036746463220