OpenAI CEO Sam Altman address 'bumpy' chatGPT-5 rollout on Reddit

During a Reddit ask-me-anything session on Friday, OpenAI CEO Sam Altman and key members of the GPT-5 team were flooded with questions about the new model and requests to bring back GPT-4o, its previous model.

One of the new and USP features that GPT-5 rolled out is a real-time router that decides which model to use for a particular prompt, either responding quickly or taking additional time to ‘think’ through answers.

But multiple people in the AMA on the Reddit complained GPT-5 wasn’t working as well for them as 4o did. According to the media reports, Altman said the reason GPT-5 seemed “dumber” was that the router wasn’t working properly when it was rolled out Thursday. He wrote, “As we mentioned, we expected some bumpiness as we roll out so many things at once. But it was a little more bumpy than we hoped for!”

“GPT-5 will seem smarter starting today. Yesterday, we had a sev and the autoswitcher was out of commission for a chunk of the day, and the result was GPT-5 seemed way dumber. Also, we are making some interventions to how the decision boundary works that should help you get the right model more often. We will make it more transparent about which model is answering a given query,” Altman added.

People on AMA still lobbied to bring back 4o for Plus subscribers. Altman responded with writing, “We are looking into letting Plus users continue to use 4o. We are trying to gather more data on the tradeoffs.”

And Altman also stated, “We are going to double rate limits for Plus users as we finish rollout."

They also asked Altman about the ‘chart crime.’

As per the reports, he was also asked about the inaccurate chart the team presented during the live presentation that quickly sparked ‘chart crime’ jokes. During GPT-5’s presentation, a flawed chart reportedly sparked jokes online. The chart misrepresented SWE-bench Verified data, exaggerating GPT-5's accuracy compared to OpenAI o3 (69.1%) and GPT-4o (30/8%), likely due to swapped or misaligned bars, as noted in OpenAI's own SWE-bench Verified documentation released on August 7, 2025, the reports noted.