Codex adds GPT-6.1 Sol Ultrafast, and a beta that drafts the next prompt
OpenAI changelog, 8 October: Ultrafast speeds GPT-6.1 Sol in Codex and ChatGPT Work for Pro at $500 and for Enterprise and Edu that turn it on. Help Center, 9 October: Composer predictions in the Codex desktop app suggest the next message for Pro users on Astra or 6.1 Sol. Tab accepts; it does not send.
Short1 minEasy
Maya Chen1 min read
- Codex
- OpenAI
- GPT-6.1
- Sol

Photo: OpenAISource page
The Codex model OpenAI is speeding up is Sol, not the Astra checkpoint it just described to Parliament as held back. On 8 October the ChatGPT and Codex changelog said Ultrafast mode speeds token generation with GPT-6.1 Sol in Codex and in ChatGPT Work. It is on Pro at $500 a month, and on eligible Enterprise and Edu plans. Enterprise access is off until a workspace owner enables it. Inference residency for this mode is the United States and Europe, defined there as the EEA plus Switzerland.
The next day’s Help Center note is about the box you type into. Composer predictions, in beta, suggest the next message from the current Codex thread. After the agent replies, a suggestion can appear in the composer. Tab accepts it. Accepting does not send; the user can edit first. The beta is for personal ChatGPT Pro users 18 and older, in supported regions, on the latest Codex desktop app, and only on local and SSH threads that are using GPT-6 Astra or GPT-6.1 Sol. It is on by default, under Show predictions in the Composer section of General settings. During the beta, generating the suggestion does not count toward Codex limits or credits. The message you actually send does. CLI 0.162.1, dated 9 October on the same changelog, is a smaller fix: a crash when an async question ran over several lines, and a startup failure when a background server’s feature flags did not match the CLI defaults.


