The anticipation round OpenAI’s GPT-5 mannequin lastly got here to an finish as the corporate launched the newest mannequin to energy ChatGPT throughout a live-stream on Thursday. OpenAI CEO Sam Altman described GPT-5 as a significant step ahead in AI and in contrast the mannequin’s skills to that of a PhD-level skilled throughout varied fields.
OpenAI says the brand new mannequin is healthier at reasoning, writing, coding, accuracy, health-related queries, and extra, whereas having considerably fewer hallucinations (making stuff up that is not true). The GPT-5 launch began rolling out to customers quickly after the launch announcement, and the largest change thus far has been bidding adieu to earlier OpenAI fashions.
OpenAI was using GPT-4o to deal with regular queries whereas o3 or o4 mini dealt with duties involving reasoning skills. Nonetheless, with the GPT-5 rollout, OpenAI has eliminated the mannequin picker choice from ChatGPT, and customers can not select which mannequin they need to use like earlier than — resulting in frustration amongst social media customers.
Notably, GPT-5 comes with an environment friendly mannequin and a reasoning mannequin, however customers don’t have to make the selection to set off the reasoning mannequin; OpenAI makes use of a router to determine which mannequin to interact for a specific question.
Netizens react to GPT-5 rollout:
Not solely had been social media customers indignant about shedding entry to their favorite fashions, however the consensus on on-line boards appears to be that the brand new GPT-5 mannequin provides shorter solutions whereas having much less character than the earlier variations. Furthermore, many customers who spent hours interacting with the chatbot now say that they’ve fewer prompts to work with — even with a ChatGPT Plus subscription.
“They’ve utterly ruined ChatGPT. It’s slower, even with out the considering mode. It has such brief replies and it will get a few of the most simple issues improper. It additionally doesn’t take heed to the directions you give and simply does no matter it needs to do.” wrote one person in a Reddit publish on the ChatGPT sub-reddit.
“They did all the pieces of their energy to shorten the solutions to be as particular and focused to save cash. They’ve eliminated the emotional intelligence of the AI as a result of if it’s not fascinating to speak to folks received’t spend all day simply chatting with it, which can save them cash. Besides it’s truly going to earn them thousands and thousands of misplaced subscriptions.” reacted one other person.
“It undoubtedly doesn’t have the identical vibe as 4o. It’s extra organized in some methods, however I discover the responses whereas correct are shorter and extra clipped in tone. ” one more person famous
“Solutions are shorter and, thus far, not any higher than earlier fashions. Mix that with extra restrictive utilization, and it looks like a downgrade branded as the brand new hotness.” opined one person
“I actually really feel like I simply watched an in depth pal die.” said a person on the elimination of GPT-4o and different fashions from ChatGPT.