OpenAI released the GPT-5.1 model; More friendly and smarter than the previous model

Three months after the disappointing release of GPT-5, OpenAI today unveiled GPT-5.1; The update that seems to be a direct attempt to resolve criticisms and restore excitement to users. OpenAI says this upgrade makes ChatGPT smarter, warmer and more enjoyable.

The launch of GPT-5 in August, despite extensive publicity, failed in practice and users protested its changes. The pressure was so great that OpenAI was forced to return the previously popular model, GPT-4o, to the platform as an opt-in option just one day after GPT-5 was released.

This failure had bigger consequences. Microsoft, OpenAI’s strategic partner and major investor, is reportedly leaning toward competing models, notably Entropic, after seeing GPT-5’s lackluster performance. Microsoft now uses entropic models in important tools such as Copilot Researcher and its GitHub Copilot.

New model GPT-5.1

GPT-5.1 was released with the aim of directly fixing these problems. This update includes two new models:

1. GPT-5.1 Instant: The most used model on ChatGPT is now warmer, more conversational, smarter and follows instructions better. OpenAI says the model surprised users with its playfulness in early tests. This model can use comparative reasoning for the first time; That is, for more challenging questions, he thinks before answering to provide more accurate answers.

2. GPT-5.1 Thinking: This advanced model for complex arguments is now more intuitive and adapts its thinking time precisely to the question: it spends more time on complex problems and answers simple questions faster.

The biggest change for users is the addition of new “personality settings” for conversation tone. The full list of character options now includes 8 items:

• Default (default)
• Professional (professional)
• Friendly (Friendly, formerly called Listener)
• Candid (honestly)
• Quirky (strange)
• Efficient (Formerly called Robot)
• Nerdy (eater)
• Cynical (cynical)

In addition, OpenAI is testing new ways to fine-tune ChatGPT’s style directly from the settings, allowing users to control things like the amount of summarization, the warmth of the tone, or even the number of emojis used. Another important improvement is that unlike in the past, any changes to personality settings are immediately applied to all your chats (even ongoing conversations).

GPT-5.1 models will be released from today first to Pro, Plus and Business users and then to free users.