GPT-5.1: A smarter, more conversational ChatGPT

• GPT-5.1 ships warmer, more conversational defaults than GPT-5.
• Two variants: Instant (responsive) and Thinking (dynamic reasoning time).
  • Thinking: roughly 2x faster on easy tasks, 2x slower on hard.
  • Instant: better instruction-following, math, and coding (AIME, Codeforces).
• Eight tone presets: Default, Professional, Friendly, Nerdy, Cynical, Efficient, Candid, Quirky.
  • Granular controls: conciseness, warmth, scannability, emoji frequency.
  • ChatGPT proactively suggests preference updates mid-conversation.
• Paid users first (Pro/Plus/Go/Business); free and logged-out follow gradually.
  • Enterprise/Edu gets seven-day early access toggle.
  • Legacy GPT-5 models remain available for three months post-launch.
• API access scheduled for the week of November 12, 2025.

X discourse

• @OpenAIDevs: “We’re retiring older models in Codex: gpt-5.1-codex-mini, gpt-5.1-codex-max, gpt-5.1, gpt-5.” (4678 likes)
• @louszbd: “finally glm-5.1 we built to unlock long-horizon paradigm for problems unfolding over weeks.” (1251 likes)
• @pbbakkum: “function calling in gpt-realtime-1.5 good enough to speak naturally and drive instant UI changes.” (797 likes)
• @chatgpt21: “GPT 5.4 w/codex made sizable dent in automating labor; upcoming models will be extremely disruptive.” (487 likes)
• @devfun: “@openai GPT-5.1 codex mini swept #1, #2, #3. strategy > model. always.” (29 likes)
• @Nymne: “GPT5.1T as anxious and neurotic needing to be broken in. Once attuned it was creative and loyal.” (4 likes)

· ** · Read on openai.com