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

Digest AI · Agents & Tools · published 2026-10-08T12:00:00Z

Canonical: https://digestai.news/story/third-moment-research-runs-10-000-agents-for-36-hours-with-gpt-6

## Summary

Third Moment Research, an AI lab, is operating 10,000 concurrent agents for autonomous research sessions lasting 36 to 48 hours. The company uses OpenAI’s GPT-6 Astra, GPT-6 Sol, GPT-6.1 Sol, and Codex to build a self-improving system that identifies agent weaknesses and generates synthetic scenarios to fix them. This infrastructure allows the team to process trillions of OpenAI tokens in a single day.

The lab applies this setup to quantitative finance, where agents test trading strategies on historical data. In one instance, Codex identified a bottleneck in backtesting US equity data, reducing processing time from approximately 3 hours to 3 to 5 minutes through caching and memory-packing optimizations. This efficiency freed up token budgets for further experiments. The team notes that newer models like GPT-6.1 Sol maintain coherent research threads over days of compute, enabling agents to investigate root causes through hundreds of parallel experiments.

Security is managed using OpenAI’s Daybreak platform, which automates penetration testing to ensure agents cannot break out of their Dockerized environments. Third Moment plans to scale this architecture to hundreds of thousands of collaborating subagents, aiming to let humans focus on optimizing the research process rather than executing individual tasks.

## Key points

- Third Moment Research runs 10,000 agents concurrently for 36-48 hour autonomous research sessions.
- Codex reduced backtesting time from approximately 3 hours to 3 to 5 minutes via optimization.
- The lab uses Daybreak to automate security testing for agent environments at scale.

## Why it matters

This demonstrates the viability of long-horizon, large-scale autonomous agent systems in high-stakes domains like finance. It highlights how combining reasoning models with specialized tools like Codex and Daybreak enables complex, multi-day investigations that were previously impossible for single agents.

## Sources

1. [How Third Moment Research runs 10,000 agents for 36+ hours](https://openai.com/index/third-moment-research) (OpenAI, 2026-10-08, primary source)

Part of the developing story: [Financial AI Agents Reach Autonomous Scale](https://digestai.news/thread/chatham-financial-uses-openai-s-gpt5-6-to-cut-trade-validation-time) (3 stories)

## Cite

Digest AI, "Third Moment Research runs 10,000 agents for 36+ hours with GPT-6", 8 October 2026, https://digestai.news/story/third-moment-research-runs-10-000-agents-for-36-hours-with-gpt-6

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/third-moment-research-runs-10-000-agents-for-36-hours-with-gpt-6.json