OpenAI Codex 'Persistent' Mode Discovered, Hinting at Long-Running Agents
imjustnewatai · x · 2026-08-27
A new 'Persistent' mode has been spotted in OpenAI's Codex settings, described as 'Continue working until put to sleep.' This mode only appears when advertised by a remote model. Codex preserves 'persistent' locally but translates it to 'disabled' for the Responses API, suggesting persistence is controlled by the agent runtime while the one-shot reasoning budget is disabled. This connects to OpenAI's June statement about its most valuable Codex work unfolding over 'hours or days' and the acquisition of Ona to provide persistent cloud environments for agents that continue working beyond the original session.
Related event: Codex Repo Reveals Unreleased 'Persistent' Reasoning Mode(2 posts)→
More from coding & agent
- Bronx high schoolers host AI dev competition to support local businesses — ziv_ravid · 2026-08-27
- Automating parsing of Indian subcontinental language books with Claude and Codex — aryaman2020 · 2026-08-27
- User claims $75 stake given to Grok bot grew to $6,140 in 48 hours — RachelVT42 · 2026-08-27
- Which coding tasks justify the highest-capability model in your workflow? — whereismy1 · 2026-08-27
- WebMCP enables seamless website integration for agents with just embedded JS — HankYeomans · 2026-08-27
- Which agent steps deserve the expensive model in long-running workflows? — Top-Construction938 · 2026-08-27