Six Frontier Models Tested in Multi-Agent Clinical POMDP
judgement_labs built a multi-agent POMDP environment where six frontier models worked through a clinical case under limited budget and an uninformed patient. HealthBench leader Qwen performed worst.
strikersam's autonomous-ai-agency trends on GitHub, executing multi-agent tasks with automatic provider failover and CEO agent orchestration, running on customer hardware with no data leaving their network.