Beginning Friday, OpenAI will cease providing access to 5 legacy ChatGPT fashions, together with the favored however controversial GPT-4o mannequin.
The 4o mannequin has been on the heart of plenty of lawsuits regarding consumer self-harm, delusional habits, and AI psychosis. It stays OpenAI’s highest scoring model for sycophancy.
Along with GPT-4o, the GPT-5, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini fashions have additionally been deprecated.
OpenAI meant to retire GPT-4o in August, when it unveiled the GPT-5 mannequin. However on the time, there was sufficient backlash for OpenAI to maintain the legacy mannequin out there for paid subscribers, who might manually select to work together with that mannequin. In a latest weblog submit, OpenAI famous that solely 0.1% of consumers have been utilizing GPT-4o, however for an organization with 800 million weekly lively customers, that small share nonetheless quantities to 800,000 folks.
1000’s of customers have rallied towards the retirement of 4o, citing their shut relationships with the mannequin.
