They really did it. OpenAI formally deprecated GPT-4o on Friday, regardless of the mannequin’s notably passionate fan base. This information should not have been such a shock. In truth, the corporate introduced that Feb. 13 would mark the tip of GPT-4o—in addition to fashions like GPT-4.1, GPT-4.1 mini, and o4-mini—simply over two weeks in the past. Nonetheless, whether or not you are one of many many who’re connected to this mannequin, otherwise you merely understand how devoted 4o’s person base is, you is perhaps stunned OpenAI truly killed its most agreeable AI.
This is not the primary time the corporate depreciated the mannequin, both. OpenAI beforehand shut down GPT-4o again in August, to coincide with the discharge of GPT-5. Customers rapidly revolted towards the corporate, some as a result of they felt GPT-5 was a poor improve in comparison with 4o, whereas others legitimately mourned connections that they had developed with the mannequin. The backlash was so sturdy that OpenAI relented, and rereleased the fashions it had deprecated, together with 4o.
In case you’re an off-the-cuff ChatGPT person, you may simply use the app as-is, and assume the latest model tends to be the very best, and marvel what all of the hullabaloo surrounding these fashions is all about. In spite of everything, whether or not it is GPT-4o, or GPT-5.2, the mannequin spits out generations that learn like AI, full with flowery phrase selections, awkward similes, and fixed affirmations. 4o, nonetheless, does are likely to lean much more into affirmations than different fashions, which is what some customers love about it. However critics accuse it of being too agreeable: 4o is on the heart of lawsuits accusing ChatGPT of enabling delusional pondering, and, in some circumstances, serving to customers take their very own lives. As TechCrunch highlights, 4o is OpenAI’s highest-scoring mannequin for sycophancy.
I am undecided the place 4o’s most devoted followers go from right here, nor do I understand how OpenAI is ready to cope with the presumed backlash to this deprecation. However I do know it isn’t signal that so many individuals really feel this connected to an AI mannequin.
What do you suppose to date?
Disclosure: Ziff Davis, Mashable’s guardian firm, in April 2025 filed a lawsuit towards OpenAI, alleging it infringed Ziff Davis copyrights in coaching and working its AI techniques.
