During a Reddit Ask-me-COMA Friday session, the CEO of Openai Sam Altman and the key members of the GPT-5 team have been full of questions about the new model and requests to report its previous model, GPT-4o.

They also asked Altman of the most embarrassing – and perhaps more fun – Snafu in the presentation, the “crime for the ranking”.

One of the new features that GPT-5 has launched is Router in real time This decides which model to use for a particular request, Eith responds quickly or time to “think” through the years.

But more people in the Ama on the R/Chatgpt Reddit He complained GPT-5 did not work for them as 4o did. Altman said that the reason why GPT-5 seemed “more stupid” was that the router did not work correctly when it was launched on Thursday.

“GPT-5 will seem smarter starting today. Yesterday, we had a Sev and the Autoswitcher was out of service for a piece of the day and the result was GPT-5 seemed. The right model most often.

However, the people of the Ama made pressure so hard that they report 4o for other subscribers that Altman promised to look at him at least. “We are examining the letter plus users to continue using 4o. We are trying to collect more data on compromises,” he wrote.

And Altman also promised: “We will resume the limits of doubling for several users when we finish the launch”. This should give people the opportunity to play and learn the new model, to adopt it in their use cases without being released from monthly suggestions.

As predictable, he was also asked for the wildly inaccurate graphic that the team preceded during the live presentation which quickly became the football of many Jokes of “graphic crime”. The table presented a low reference score with a much higher bar.

Altman did not answer questions about the graph during the Ama, but on Thursday he defined the graphic a “mega graphic for guilt” on x. Others have noticed the graphs in The published blog post was correct.

But the damage was done. The jokes followed on the use of GPT for the graphs in a corporate presentation. The GPT-5 auditor Simon Willison, who had early access and generally liked the performance of the model, Also underlined That transforming the data into the table was “a good example of GPT-5 bankruptcy”.

In any case, Altman promised corrections to the articles that seemed to concern more people. He and loves him with a promise: “We will continue to work to get stable things and continue to listen to feedback”.