GPT Advent

Shipped · day 19 of 28

Faster dot calls, cache-preserving reasoning changes and CLI updates

Dot calls now connect much faster, and changing reasoning settings during a Codex or Work conversation no longer invalidates the cache. The Codex CLI adds reusable prompt presets, inline model and permission settings, enterprise MCP sign-in and Windows corporate-proxy authentication, plus fixes for failures when resuming long sessions.

Did you love it?

Back to the calendar