ChatGPT Is Less Flattering and More Concise Now, If You Are on Free or Go

If free ChatGPT has felt blunter since the end of October, that is a change OpenAI made on purpose and did not tell you about in the app. On October 29, 2026 it updated the model behind ChatGPT's chat tab for free users, and the entry describing it sits on a page aimed at developers.

Key Takeaways

• The change is two things: "more concise responses" and "reduce overly flattering language." Those are OpenAI's words.
• It applies to Free and Go users, and to people using ChatGPT without signing in. That is the large majority of people who use it.
• Plus and Pro are mostly untouched. OpenAI says the main GPT-6 model on those plans is unchanged, and the new behaviour appears only on the fallback model you get after hitting your message limits.
• There is nothing to switch on, and nothing to switch off. It is a change to the model, not a preference.
• OpenAI published it on its model release notes, not the ChatGPT release notes. The consumer page, read live on November 2, 2026, does not mention it.
• An updated Model Spec landed the same day and contains the principle behind it: the model "should not infer that preference merely from what it thinks they want to hear."
• OpenAI went the other way 14 months ago. In August 2025 it made GPT-5's default personality "warmer and more familiar" after complaints it was too reserved.

Who Gets It

Almost everyone who does not pay, and almost nobody who does. This is the whole of OpenAI's statement, from the October 29, 2026 entry on its model release notes page, read live on November 2, 2026:

We're updating GPT-6 in Chat to give more concise responses and reduce overly flattering language. This update applies to Free and Go users, and people using ChatGPT without signing in. On Plus and Pro, it applies only to the model ChatGPT can switch to after you reach GPT-6 message limits. The main GPT-6 model on Plus and Pro is unchanged.

The Plus and Pro line is worth reading twice, because it describes a tier of ChatGPT most subscribers forget exists: when you exhaust your GPT-6 allowance, ChatGPT keeps answering on a lesser model. That fallback is now the concise, less flattering one. So paying readers will meet this change at the end of a heavy day rather than never.

What "Less Flattering" Actually Means

It means fewer compliments at the top of answers, and OpenAI has a word for the thing it is removing. In August 2025, announcing the opposite change, it defined the terms itself on the same release notes page:

Warmth here means small acknowledgements that make interactions feel more personable -- for example, "Good question," "Great start," or briefly recognizing the user's circumstances when relevant. This is different from sycophancy, which we define as excessive flattery that can feel insincere.

That is the clearest published description of what is being dialled down. "Good question" is the shape of it.

What OpenAI does not publish is how much. There is no measurement in the entry, no before-and-after example, no target length, and no evaluation cited. The change is real and its size is not something anyone can honestly quote, including us.

The Model Spec Rewrite Behind It

OpenAI published an updated Model Spec the same day, and one sentence in it is the principle the model change enforces. The Spec is OpenAI's public description of how it wants its models to behave, and the revision reorganizes it around five values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care.

The line that matters for tone is in the section on what was retired:

At the same time, we're retiring most of the Spec's detailed style guidance. Style should adapt to what people prefer [...] The Spec retains stronger defaults, such as approaching questions objectively. Users and developers can explicitly ask the model to take a particular perspective, but the model should not infer that preference merely from what it thinks they want to hear.

In plain terms: ChatGPT is meant to stop guessing which answer will please you. Two other changes in the same revision are worth knowing because they will show up in later coverage: OpenAI "adjusts some behavioral defaults", and it now calls the model "GPT" rather than "the assistant" throughout the document. It also says plainly that "our models do not yet fully reflect it", so the Spec is a target rather than a description.

What You Can Do About It

Ask for the old style, because the controls are still there.

OpenAI has shipped tone presets and personalization controls since late 2025, and this change does not remove them. What it removes is the default. If you liked being told "great question", you now have to ask for it.

One thing not to do: do not read a blunter answer as a worse answer. OpenAI's stated aim is concision, and the same release notes page has described tighter answers as an improvement for two years running. Whether you agree is a matter of taste, and taste is the thing OpenAI just said it would stop guessing at.

Why You Did Not See This Announced

Because it was filed on the page consumers do not read. There are two OpenAI release notes pages, and the difference matters:

We read both in full on November 2, 2026. The consumer page's newest entry is October 30, 2026, account switching on iOS and Android, and a search of the whole document finds no 2026 entry about concision or flattery. The model page carries the change as its second-newest entry.

This is a pattern rather than an accident. A model's personality is the single most noticeable thing about it to someone who does not read release notes at all, and it is documented in the place built for people who do.

The Bottom Line

If you use ChatGPT free, your assistant got terser and stopped buttering you up, as of October 29, 2026. Nothing is broken and nothing needs doing. If you miss the old tone, set it in Personalization.

If you pay for Plus or Pro, nothing changed until you run out of GPT-6 messages, and then the model you fall back to is the new one.

If you are comparing chatbots, this narrows one of ChatGPT's soft advantages. For how it stacks up against Gemini on everyday writing, see ChatGPT vs Gemini for everyday use, and for what free tiers include right now, the best free AI chatbot.