OpenAI returns old models to ChatGPT as Sam Altman admits ‘bumpy’ GPT-5 rollout
1 min read
Summary
OpenAI’s new flagship Generative Pre-trained Transformer (GPT) model, GPT-5, experienced a problematic launch, with CEO Sam Altman citing issues including faulty model switching, poor performance and user confusion.
In response, OpenAI has partially walked back some of the platform changes and given users access to the earlier GPT-4o model, and promised changes including updating the UI to make it clearer to users which model is responding to a query.
Some beta testers were also disappointed that OpenAI unilaterally upgraded their ChatGPT experiences to GPT-5 and took away access to older models, and that real-world performance of GPT-5 lags behind its hype.
This has given rivals an opportunity to gain ground, and put pressure on OpenAI to prove that GPT-5 is more than just an incremental update.