OpenAI’s internal model Garlic, built on improved pretraining and bug fixes, may launch as GPT-5.2 or GPT-5.5 early next year.
Recent reports indicate that OpenAI is advancing a new model internally codenamed Garlic, which appears to be the centerpiece of its near-term strategy to improve ChatGPT. CEO Sam Altman reportedly declared a “code red” to boost ChatGPT’s performance, and Garlic is now seen as the likely vehicle for that upgrade, potentially launching as GPT-5.2 or GPT-5.5 in early 2026. Garlic is distinct from another model under development, Shallotpeat, but it benefits from improved pretraining methods and bug fixes discovered during that work.
Source: radicaldatascience.wordpress.com