OpenAI 启动 28 天 Codex 升级计划：首日 GPT-6 Astra / GPT-6.1 Sol 默认提速约 50%
OpenAI pledged one user-noticeable Codex or ChatGPT Work improvement per day for 28 days; day one (Oct 6) shipped about 50% faster default GPT-6 Astra and GPT-6.1 Sol via server scheduling, with OpenCode, Pi, Amp and Devin (Sign in With ChatGPT) gaining the same speedup.
On October 5, OpenAI announced a 28-day Codex and ChatGPT Work improvement streak, shipping one user-noticeable upgrade per day. Day one shipped a server scheduling optimization that lifted default GPT-6 Astra and GPT-6.1 Sol response speed by roughly 50 percent, with no user action required and effect rolling in within about two hours. The speedup comes from how the servers handle requests, not from changes to model reasoning itself, so per-step quality is preserved.
The improvements remain in effect for the same users; outside tools that use Sign in With ChatGPT, including OpenCode, Pi, Amp and Devin, get the same speedup. OpenAI also committed to fully resetting usage limits on days when no improvement ships.
This is OpenAI's first explicit "ship daily" commitment since DevDay 2026. The 28-day streak is aimed at cutting cumulative latency in repeated workflows, particularly for coding agents that loop through write, test, fix. The streak sits alongside GitHub Copilot's October 1 computer-use desktop preview and OpenCode's late-September Operation Cheepseek Phase 2 (DeepSeek V4.1 Flash $60 permanent quota) as the spine of the October coding-agent month.