NAC Agent Framework Revealed: Powers Massive Chunk of Pre-training Code
MaziyarPanahi · x · 2026-08-14
An open-source model team announced the introduction of NAC, an internal agent harness, alongside the beta of their expanded Open Models API.
NAC has been used daily by the research team since April and is built exclusively for long-running, asynchronous, and hands-off tasks. Notably, a significant portion of the code committed to their pre-training, post-training, and data pipelines over the last three months has been powered by NAC, often driven remotely.
Related event: Arcee AI Open-Sources NAC Framework for Long-Running Agent Tasks(10 posts)→
More from coding & agent
- Customer Support's Future: Multi-Tier AI Agents Escalating to Solve Issues — round · 2026-08-14
- Stanford's CooperBench: Multi-Agent Cooperation Fails More Than Solo Agents — _Hao_Zhu · 2026-08-14
- Factory AI Launches Agent Effectiveness to Track AI Spend ROI — matanSF · 2026-08-14
- Mendel Gödel Machine: Recursive Self-Improving Agents via Comparative Evolution — burny_tech · 2026-08-14
- Coinbase CEO: Adopting AI Requires Reworking Old Habits — J0se · 2026-08-14
- Claude Code Ports 210K Lines of 1990s C++ to Web in Two Months — moenig · 2026-08-14