Back in August, OpenAI unveiled GPT-5, a major frontier model update to ChatGPT. Three months later, OpenAI is back with GPT-5.1, which it says is “now warmer, more intelligent, and better at following your instructions.”
Instant and Thinking improvements
According to OpenAI, GPT-5.1 focuses on both communication style and intelligence. The company summarizes improvements to both the instant model and thinking model:
GPT‑5.1 Instant: our most-used model, now warmer, more intelligent, and better at following your instructions. GPT‑5.1 Thinking: our advanced reasoning model, now easier to understand and faster on simple tasks, more persistent on complex ones.
Based on provided examples, GPT-5.1 instant responses read less like templated data responses and more like natural writing. OpenAI says the warmer, more conversational tone is the new default.
However, the company is also adding “more intuitive and effective controls so ChatGPT can better match the tone you want in responses.” Personality options include default, friendly, efficient, professional, candid, and quirky.
Following specific instructions is another improvement to the instant model in GPT-5.1. In the example provided, OpenAI shows the model correctly responding to prompts with the instructions to only respond with six words. If it works as advertised, this will be a nice change to one of the more frustrating flaws of ChatGPT before now.
GPT-5.1 Thinking, on the other hand, has new tuning that makes it faster on quick tasks and slower on tasks that should require more consideration. It also inherits the warmer tone by default.
Like with GPT-5, the auto-routing feature will ping which model is best for each task by default, or you can select which model you want to use per prompt.
Availability
... continue reading