OpenAI eases the rate limits on GPT-5 and commits to enhancing its personality for improved SEO performance.

OpenAI is gradually addressing various concerns surrounding GPT-5, particularly regarding rate limits and its personality, which has faced criticism for being less affirmative. In a recent support document, OpenAI confirmed the restoration of older models for paid customers, allowing access to GPT-4o, GPT-3, and others. Users can easily select these legacy models through the model selector. Furthermore, GPT-5 features an automatic switch between Fast and Thinking modes, with additional options available directly from the model picker. For instance, GPT-5 Fast provides instant answers, while the Thinking mode offers deeper reasoning capabilities. The GPT-5 Thinking limit has also been increased to 3,000 per week for GPT Plus subscribers.

In a post on X, OpenAI CEO Sam Altman announced that the company is working on an update to enhance GPT-5’s personality. This update aims to make GPT-5 feel warmer than its current iteration, although it may not reach the same level of affirmation as GPT-4o. OpenAI is also exploring the possibility of per-user customisation of model personality to better meet individual preferences. As the landscape of AI continues to evolve, these updates reflect OpenAI’s commitment to improving user experience and addressing feedback effectively. 

Categories: Model Features, Personality Updates, User Customization 

Tags: OpenAI, GPT-5, Rate Limits, Personality, Legacy Models, Model Selector, Fast, Thinking, Customization, Subscribers