Key Highlights:
- GPT-5.2 is now expected to launch December 11, based on shifting Polymarket predictions.
- Notion is internally testing an unreleased model called Olive Oil Cake, which is believed to be GPT-5.2.
- All new details point towards an imminent release, though OpenAI hasn’t officially confirmed anything yet.
Since last week, there has been growing chatter on the internet about OpenAI’s upcoming GPT-5.2 model. A recent report from The Verge hinted at December 9 as the possible release date of the new model. But, OpenAI didn’t release the model yesterday. No doubt, you must be wondering when exactly OpenAI will launch the GPT-5.2 model, right?
Polymarket activity & market sentiment points at December 11 as GPT-5.2 release date
Well, thanks to folks at Testing Catalog, we may have some information on the release date. Per the report, Polymarket activity and current market sentiment points to Thursday, November 11 as the potential release date of the GPT-5.2 model.
That being said, there’s no word from OpenAI on the official release date of the GPT-5.2. Reports from last week suggested that OpenAI is fast-tracking GPT-5.2 as part of Sam Altman’s company-wide “code red” call after Google pushed Gemini 3 last month, which beats OpenAI’s model on several benchmarks.
All that said, the most recent reference to the GPT-5.2 model comes from Notion. For the uninitiated, the platform has historically tested OpenAI’s unreleased models ahead of launch. As reported, a new option, internally called “Olive Oil Cake,” has appeared for some testers. Interestingly, it comes with the OpenAI logo and a unique identifier that is different from GPT-5.1.
Well, folks at TestingCatalog believe that it is GPT-5.2. As of now, we don’t know whether the model is on par or exceeds the performance of Gemini 3. But, with December 11 rumored as the GPT-5.2 release date, we are not far from getting to know its capabilities.
New, rumored image models from OpenAI & Google
Not to forget, OpenAI is also reportedly testing new AI image models which are internally codenamed “Chestnut” and “Hazelnut.” Early testing reports suggest that the quality of image generated by the rumored new models are far from Nano Banana Pro-generated images. Meanwhile, Google is also reportedly testing an affordable image model called Nano Banana 2 Flash, which is rumored to be powered by Gemini 3 Flash.
Are you excited for new models? If yes, which one are you most excited for? Is it Google’s or OpenAI’s? Drop your thoughts in the comment below.
