In a separate Substack post, OpenAI CEO for Applications, Fidji Simo, said these chat models are “trained using the same stack as our reasoning models,” adding that they “score higher on factuality and complex problem-solving than GPT-5, while also introducing a more natural, conversational tone.”
The update expands ChatGPT’s personalization controls with presets such as professional, friendly, candid, quirky, efficient, cynical, and nerdy, and aims to improve the reliability of custom instructions so they persist across multi-turn conversations.
Implications for enterprises
Analysts point out that the real story with GPT-5.1 is not the headline features but the subtle behavioral shifts that remove a lot of the friction enterprises quietly learned to live with this past year.
This story originally appeared on Computerworld
