OpenAI’s newest flagship model, GPT-5, was introduced as an all-in-one AI system designed to simplify ChatGPT by removing the need for a model picker. However, just days after launch, the model picker has returned, and with more options than before.
When GPT-5 launched last week, OpenAI promoted its built-in “router” that would automatically choose the best way to answer a user’s question. This was meant to replace the manual selection of models, a feature CEO Sam Altman had previously criticized. But now, users can pick between three GPT-5 modes: Auto, Fast, and Thinking. Auto uses the original routing system, while Fast and Thinking give direct control over speed and depth of responses.
In another unexpected move, OpenAI has brought back older AI models after strong user backlash. Paid subscribers can now access GPT-4o, GPT-4.1, and o3, with GPT-4o even set as the default. These models had only been removed last week, sparking criticism from loyal users. Altman has promised that if GPT-4o is ever retired again, users will be given plenty of notice.
OpenAI is also working to adjust GPT-5’s personality to make it warmer and allow more customization so users can set their preferred tone and style. The launch, however, was not without issues. On the first day, the GPT-5 router malfunctioned, leading to performance complaints. Altman addressed the problem during a Reddit AMA, while OpenAI’s VP of ChatGPT, Nick Turley, praised the team’s quick fixes.
The changes highlight a challenge for OpenAI: while faster and smarter AI is important, many users form strong attachments to specific models because of their tone, verbosity, and personality. Balancing new technology with user preferences is proving to be more complex than expected.
