How Third Moment Research runs 10,000 agents for 36+ hours
Third Moment Research is running 10,000 agents continuously for more than 36 hours, according to OpenAI News. The lab's setup draws on GPT-6 Astra and GPT-6.1 Sol, Codex, and Daybreak for cybersecurity.
The headline emphasizes duration: 10,000 agents sustained for 36+ hours. The excerpt describes agents that focus for days and improve with use, with Daybreak positioned for cybersecurity work.
Details on architecture, cost, benchmarks, and availability are not part of the announcement.
Why it matters: Long-running, self-improving agent fleets could change how developers plan for state management, cost control, and security. Teams building with these models may need to treat persistence and oversight as first-class concerns, though specifics remain unstated.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
With GPT-6 Astra and GPT-6.1 Sol, Codex, and Daybreak for cybersecurity, this AI lab builds agents that focus for days and improve with use.