$TREE With GPT-5, OpenAI has eliminated eight ChatGPT models at once. You will only be able to use GPT-4o under one condition. 🚨

After user complaints, Sam Altman has raised his hand allowing the choice of the "old" GPT-4o under certain conditions and for the moment.

The arrival of GPT-5 has brought an unexpected side effect for those of us who use ChatGPT in our day-to-day: OpenAI has eliminated eight AI models in one go, GPT-4o, GPT-4.1, GPT-4.5, GPT-4.1-mini, o4-mini, o4-mini-high, o3, and o3-pro, so that all activity is directed towards GPT-5 and its versions.

And that means a lot of changes and a lot of faith in their new model. But as the CEO of OpenAI has stated, the launch of GPT-5 was a bit more eventful than expected.

After the launch, the top executive of the ChatGPT company stood in front of users in an "Ask Me Anything" on Reddit, where they expressed their dissatisfaction with the void left by the inexplicable absence of the old and (more or less) reliable models.

OpenAI backs down with GPT-4o but with conditions.

When OpenAI launched GPT-5, it promised the smartest and fastest model ever released by the company with a strategic tool under its mandate: a simplified system to choose which version is optimal for the user's request.

It should be noted that GPT-4o had not completely disappeared: it remained operational in ChatGPT's voice mode and in Apple Intelligence, although in a much more limited form.

But the migration to GPT-5 was automatic: if you had a chat in GPT-4o, 4.1, 4.5, or their mini versions, it is now operational in GPT-5. And if you were using o3, it will now continue with GPT-5 Thinking, just like the o3-pro models, which are now transferred to GPT-5 Thinking Pro.

A user recounted that 'For months I was perfectly in tune switching between o3, o3-pro, 4.5, and 4o, depending on the task. I knew exactly what each model could offer. Now they are gone, and I have to readjust to GPT-5'.