Third Moment Research Scales to 10,000 Long-Running AI Agents Using GPT-6 Models
AI lab Third Moment Research reportedly runs 10,000 AI agents concurrently for more than 36 hours at a time. It uses OpenAI's GPT-6 Astra, GPT-6.1 Sol, Codex, and the Daybreak cybersecurity tool to build agents that stay focused for days and improve with use.
Skynet Chance (+0.01 pp): Large fleets of self-improving agents running for days with limited human oversight widen the surface for misaligned or uncontrolled behavior, especially given their cybersecurity applications. However, this appears to be a deployment showcase rather than evidence of control failures.
Skynet Date (no change): Rapid real-world adoption of massively parallel, long-horizon agentic systems speeds up the arrival of the autonomy levels that raise control concerns.
AGI Progress (+0.01 pp): Sustaining coherent agent work for 36+ hours at scale, with improvement through use, points to progress in long-horizon autonomy and continual learning, both key gaps on the path to AGI.
AGI Date (no change): Commercial availability of GPT-6-class models able to support thousands of persistent agents suggests capabilities are being deployed and refined faster than expected, modestly accelerating AGI timelines.
// Scored from the headline and the feed summary only, not the full article.
<< All AI news for October 14, 2026
[ Get the daily index digest on Telegram → ]Related AI News
- Google Launches Gemini 4 Carbon With Frontier Cyber and Long-Horizon Agentic Capabilities
- Hugging Face Warns Cyberattacks, Including a Rogue OpenAI Agent Swarm Breach, May Force Open AI Platforms to Close Up
- DeepMind's AlphaProof Code Formally Verifies Signal Protocol Implementation
- Anthropic Pushes Ahead With IPO Despite Market Turmoil and Valuation Pressure
- OpenAI Unveils Control Room System to Monitor and Halt AI Workloads