How Third Moment Research runs 10,000 agents for 36+ hours

via OpenAI · 3 views · 1d ago

Imagine AI that doesn't lose focus. Third Moment Research is now running 10,000 agents for over 36 hours straight. They're leveraging advanced models like GPT-6 Astra, GPT-6.1 Sol, and Codex, along with Daybreak for cybersecurity tasks. This isn't just about quick processing; these agents are engineered to maintain concentration for days, constantly improving with use. This breakthrough means AI can tackle complex, long-term problems with unprecedented persistence and autonomy. It's a significant leap towards truly effective, self-improving AI systems.

← Back to the feed