During a Reddit AMA on Friday, OpenAI CEO Sam Altman and the GPT-5 team answered questions about the new model and requests to bring back GPT-4o.
One funny moment people asked about was a mistake in a presentation, nicknamed the chart crime.
GPT-5 has a new feature called a real-time router, which chooses the best model for each question, either fast or more thoughtful. But on Thursday, a technical issue (“sev”) broke the router, so GPT-5 sometimes gave worse answers.
Altman said the problem is now fixed, and GPT-5 should feel smarter. He also said they’ll improve how the router picks models and make it clearer which model is answering.
Many users still want GPT-4o back. Altman said they’re considering letting Plus subscribers use it again and are collecting data to see the pros and cons.
He also promised to “double the rate limits for Plus users” once the rollout is complete, so people can experiment with GPT-5 without running out of prompts.
As expected, he was asked about the inaccurate chart shown during the live presentation, which became the target of many “chart crime” jokes. The chart showed a lower score but used a much taller bar.
Altman didn’t talk about the chart issue during the AMA, but on Thursday, he called it a “mega chart screwup” on X. The charts in the blog post were correct, but the mistake in the presentation had already sparked jokes about using GPT for corporate charts.
GPT-5 reviewer Simon Willison, who had early access and mostly liked the model, said that making a data table was a clear example of GPT-5 failing.
Altman promised to fix the main issues people were worried about and ended the AMA by saying, “We will keep working to make things stable and will keep listening to feedback.”