ChatGPT has undergone another significant enhancement, this time with the introduction of GPT-5.1. OpenAI asserts that the update will enable “easier customization” of the chatbot while offering a “smarter, more conversational” interaction overall.
Today, the company announced the rollout of GPT-5.1 in a [blog post](https://openai.com/index/gpt-5-1/), highlighting the inclusion of two new models named GPT-5.1 Instant and GPT-5.1 Thinking. Instant aims to deliver a “warmer, more intelligent” model that excels in adhering to user instructions, whereas Thinking will concentrate on providing sophisticated reasoning. OpenAI indicates that ChatGPT will not only manage simple tasks more effectively but will also possess an improved understanding of intricate tasks.
The company shared examples illustrating how the outputs from GPT-5 and GPT-5.1 Instant will vary, showing that the company’s idea of “playfulness” is indeed more apparent in the communications. Moreover, the firm mentions that GPT-5.1 Instant will employ adaptive reasoning to assess whether it should contemplate before answering more challenging inquiries, enhancing accuracy without significantly slowing down the system.
New avenues for customizing ChatGPT
Instant isn’t the sole model designed to deliver a warmer interaction, as OpenAI states that Thinking will also provide a more comforting and empathetic response mechanism, even highlighting the distinction between it and GPT-5. However, the notable point is that this “release is a significant advancement” for the models, not just in functionality but also in user-friendliness. A considerable aspect of this is attributed to the newly introduced personalization features.
In addition to adding support for GPT-5.1, users can now further tailor ChatGPT’s responses based on personality traits. The examples shared in OpenAI’s blog post present options like Candid, Quirky, Efficient, Professional, Friendly, and others. Once chosen, the personality will apply across all AI models. The prior Nerdy and Cynical options will still be available.
The company is also experimenting with more detailed customization methods, which may allow for even finer adjustment options in upcoming releases similar to the [custom instructions](https://www.bgr.com/tech/chatgpt-custom-instructions-were-only-for-paid-subscribers-but-now-you-can-use-them-for-free/) currently available for the AI. There was no mention of the recently [leaked ChatGPT group chat feature](https://www.bgr.com/2024447/chatgpt-group-chat-feature-leaked/), so we will have to wait for further details on that front. Auto will continue to enable you to depend on the AI to identify which model is most appropriate for a given task. Additionally, if the more inviting response from GPT-5.1 is excessive, you can always revert to GPT-5 through the legacy models option in ChatGPT, just as you [used GPT-4o when GPT-5 launched](https://www.bgr.com/1936724/use-gpt4o-chatgpt-hate-gpt5-upgrade/).
