Openai’s CEO, Sam Altman, overpromtered on GPT-5, and real-life results have decreased, but it seems that a new update is rolling that may address some concerns.
GPT-5 is a state-of-the-art model. In our trials, Bleepingcomputer found that GPT-5 does actually performs well in coding. This was much faster than other Openai models, including O3.
However, GPT-5 struggles to be ‘creative’ in writing, and when expecting users it often fails to switch to its new logic abilities.
At its top, we have seen that GPT-5 often produces small materials when it is clearly asked to give more details.
Some believe that GPT-5 is throtting tokens to reduce the cost, but Openai’s CEO, Sam Altman, argues that a bug has caused unexpected problems with GPT-5.
“Yesterday, the autosquelil broke and was out of the commission for a part of the day, and the result was GPT -5, it was Dumber,” Sam Altman wrote In a post on X.
He also said that Openai GPT Plus will double the GPT-5 rate limit for users, but it will take a few days.
“We will allow plus users to choose to continue using 4Os. We will see the use because we think how long to offer heritage models,” Sam said.
The GPT-5 will also be smarter starting today, and Openai is testing a new togle that will force GPT-5 to use logic abilities.
“We will make it more transparent that which model is responding to a given query. We will replace the UI to make it easier to trigger thinking manually.”
GPT-5 is still rolling out for payment and free users, and it can be some time before receiving all.
Meanwhile, Openai is planning to restore the old model, including 4Os for plus customers.
Right now, you can only use these heritage models when you are paying $ 200 for pro subscriptions.
Malware targeting password stores increased 3x as the attackers secretly carried out the perfect history landscape, infiltrated and exploited important systems.
Search for the top 10 Metter Att & CK techniques behind the 93% attacks and how to defend them.
