The market for the third-best AI agent lab by end of September shows tight clustering around 40–50% implied probability across numerous contenders, reflecting fragmented trader views on a rapidly evolving field. OpenAI, Anthropic, Moonshot, and others compete on demonstrated agentic performance—such as autonomous task completion, tool use, and multi-step reasoning in large language models—rather than raw model scale alone. Key swing factors include imminent capability benchmarks, developer conference demos, or partnership announcements that could shift relative positioning before the resolution window closes. Limited major releases in the past month have kept outcomes closely matched, with traders weighing each lab’s track record against potential late-cycle improvements in safety, integration, or real-world deployment metrics.