OpenAI is refreshing the ChatGPT experience with a model upgrade and new personalization controls that land today.
November 12, 2025 — OpenAI has released GPT‑5.1, an update to the GPT‑5 series that pairs a friendlier, more natural ChatGPT with clearer reasoning. The rollout includes two models—GPT‑5.1 Instant and GPT‑5.1 Thinking—and begins today, with availability expanding over the next few days.
What’s actually new in the models
GPT‑5.1 Instant now uses adaptive reasoning to decide when to “think” before answering harder questions, improves instruction following, and aims for a warmer baseline tone. OpenAI also reports stronger performance on math and coding benchmarks such as AIME 2025 and Codeforces. GPT‑5.1 Thinking varies its reasoning time by task complexity and writes with less jargon for easier-to-follow explanations. OpenAI
Personalization: pick a personality (and tune it)
OpenAI is expanding ChatGPT personality presets. Professional, Candid, and Quirky join the refreshed Default, Friendly (formerly Listener), and Efficient (formerly Robot); Nerdy and Cynical remain. Beyond presets, OpenAI is experimenting with granular controls—how concise, warm, or scannable responses are, plus emoji frequency—and ChatGPT can suggest updating these preferences mid‑conversation. Updates now apply across all chats instantly, not just new ones.
Rollout, API names, and the sunset window
The update starts today for paid tiers (Pro, Plus, Go, Business) and then extends to free and logged‑out users. Enterprise and Edu receive a seven‑day early‑access toggle (off by default). For developers: GPT‑5.1 Instant arrives in the API as gpt‑5.1‑chat‑latest, and GPT‑5.1 Thinking as GPT‑5.1 later this week, both with adaptive reasoning. OpenAI will keep GPT‑5 available in ChatGPT’s legacy dropdown for three months while it transitions, and it has posted a system card addendum detailing safety approach changes.
Why OpenAI is recalibrating the ChatGPT experience
VentureBeat says GPT‑5 drew mixed reviews, so OpenAI is rebooting ChatGPT around GPT‑5.1 while leaving older models available for comparison.
Axios notes OpenAI is chasing a friendlier tone while warning about unhealthy attachments, and says the company is working with outside experts.
What it means for users and teams
You’ll feel a tone shift. Both 5.1 models aim to be more conversational and empathetic, especially on everyday prompts where warmth matters, while still tightening instruction‑following for task accuracy.
You’ll spend less time fiddling with settings. The expanded preset picker (plus on‑the‑fly suggestions) makes it faster to land on a response style that suits the moment—polished for work, candid for coaching, or quirky for brainstorming.
Reasoning should feel more “right‑sized.” Instant stays fast but now thinks longer when needed; Thinking allocates more time to hard problems and speeds through easy ones. Expect fewer overthought replies on simple requests and more persistent reasoning on complex ones.
Change management is smoother. Keeping GPT‑5 in the legacy menu for three months gives teams a comparison window and a safer path to standardize on 5.1 without breaking workflows.
How to try GPT‑5.1 now
- In ChatGPT: Check the model picker for GPT‑5.1 (rollout is staged). Explore Personalization in settings to choose a base style and tweak response qualities; changes now apply across existing chats.
- On the API: Target
gpt‑5.1‑chat‑latestfor Instant and GPT‑5.1 for Thinking when they go live later this week. If you’re on enterprise plans, watch for the early‑access toggle in the admin controls.
Key takeaways
- GPT‑5.1 refreshes the default ChatGPT experience with warmer conversation and clearer reasoning in two flavors—Instant and Thinking—starting today.
- New personality presets (Professional, Candid, Quirky, etc.) plus granular style controls make it easier to match tone to task; settings now update across all chats in real time.
- Rollout begins with paid users and expands to free; API names:
gpt‑5.1‑chat‑latest(Instant) and GPT‑5.1 (Thinking). GPT‑5 remains available for three months as a fallback. - The shift also course‑corrects after mixed reception to GPT‑5, without removing older models immediately.
The bottom line
OpenAI is pairing incremental IQ gains with deliberate EQ tuning. If GPT‑5 felt uneven in tone or instruction‑following, GPT‑5.1 is designed to be the “daily driver” that talks more like you—without giving up the heavy‑duty reasoning headroom when you need it. OpenAI
