OpenAI has quietly retired GPT-5.2. As of June 12, GPT-5.2 Instant, GPT-5.2 Thinking, and GPT-5.2 Pro are no longer available in ChatGPT. Existing conversations that were running on GPT-5.2 are automatically redirected to the corresponding GPT-5.5 model.
What’s changing
For regular ChatGPT users: practically nothing. Conversations simply continue on GPT-5.5, which is the more capable model anyway. GPT-5.5 was released on April 23, 2026, and has since surpassed GPT-5.2 on most benchmarks.
For API developers, it’s a different story. If you’re still using GPT-5.2 model strings in your codebase, you need to migrate now. The old endpoints no longer return responses.
The bigger picture
The timing is interesting: OpenAI is cleaning up its model portfolio while GPT-5.6 is reportedly already floating through developer channels. A checkpoint codenamed ‘Kindle-Alpha’ has surfaced in testing environments — with reports of a 1.5 million token context window and improved reasoning. OpenAI has neither confirmed nor denied this officially.
At the same time, OpenAI is preparing for its IPO. The confidential S-1 was filed with the SEC on June 8, with a target valuation of roughly one trillion dollars. Cleaning up the product portfolio and removing old models fits that narrative.
Parallels with Anthropic
There’s a notable parallel here: on the same weekend, Anthropic is retiring its old Claude models, and OpenAI just did the same with GPT-5.2. Both companies are consolidating their model landscapes — fewer versions, clearer upgrade paths. For developers, the takeaway is clear: regular migrations are becoming the new normal.
Sources: