Microsoft has made the GPT-5.2 model series from OpenAI generally available in Microsoft Foundry, Microsoft said in a blog post on 11 December 2025.
Two models are available: GPT-5.2, described as the most advanced reasoning model in the series, and GPT-5.2-Chat, a more efficient model for everyday work such as information seeking, how-to guides, technical writing and translation.
Microsoft said GPT-5.2 can produce design documents, runnable code, unit tests and deployment scripts with fewer iterations. It is built on new architecture and trained on the GPT-5.1 dataset, with improved safety and integrations.
The model is optimised for agent scenarios in Foundry. Microsoft highlighted multi-step reasoning, planning across large inputs such as codebases and meeting notes, and end-to-end task coordination across design, implementation, testing and deployment.
Enterprise controls include managed identities and policy enforcement. Microsoft listed financial services, healthcare, manufacturing and customer support among the target industries.