Just three mohs after unveiling the groundbreaking GPT-5 language model, OpenAI unveiled its improved version, GPT-5.1 Artificial Ielligence. Launched today for ChatGPT subscription users, the update is billed as “friendlier, smarter and better at following directions” and aims to enhance the user experience in two key areas of ielligence and communication style. This new version includes two optimized models named “Insta” and “Thinking”, each designed for differe tasks.
The GPT-5.1 Insta model, which is the most widely used model of this platform, now has a more natural and warm tone, and its responses are less like machine data and more like human writing. OpenAI has chosen this friendly tone as the new default mode. However, for the first time, iuitive and effective corols have been added to personalize the tone of responses. Users can choose their desired personality from options such as default, friendly, efficie, professional, blu and weird. Another importa developme of this model is its ability to strictly follow specific instructions; For example, if it is asked to summarize an answer in six words, it will now do so correctly.
Another model, GPT-5.1 Thinking, designed for advanced reasoning, now has new optimizations. This model works faster in performing simple tasks and shows more stability and persistence in complex issues that require more coemplation. This model also takes advaage of the new warmer tone by default. The automatic routing system will still select the best model for each request, but users will also have the option to manually select the model they wa. This update will be activated for all paid users (Pro, Plus, Go and Business) in the next few days, and then it will be available to free users. Due to negative user feedback following the iroduction of GPT-5, OpenAI has announced that the older model will remain available to paid subscribers for at least another three mohs.
Source: 9To5Mac
