Third Moment Research runs 10,000 agents 36+ hours with GPT-6 models

third moment research successfully ran 10,000 agents for 36+ hours using gpt-6 astra and gpt-6.1 sol, alongside codex and daybreak for cybersecurity applications. these agents demonstrated enhanced focus and performance over extended periods. this development has implications for ai labs and industries relying on prolonged agent deployment. the use of gpt-6 models and other technologies enabled this achievement. businesses and researchers interested in long-term ai agent operations should consider the potential benefits and applications.