Z.ai rolled out GLM‑5.2, its latest coding‑focused model, granting all subscribers a usable 1‑million‑token context window. The release targets developers building long‑horizon agentic applications and arrives ahead of an API launch scheduled for next week. Existing Lite, Pro, Max and Team plans receive immediate access via the company’s devpack documentation. It also keeps the series' edge in producing production‑grade code.

Z.ai promises an open‑source drop under the MIT License next week, though exact timing remains vague. The move follows growing demand for frontier models that can retain context over massive token spans, a niche where GLM‑5.2 claims a lead. By coupling extensive context with proven coding performance, the model aims to streamline complex code generation pipelines. Developers can fine‑tune it locally after weight release.

Early adopters report faster iteration on large codebases and smoother integration with tools like Claude Code and Cline. With the API and chatbot endpoints slated for release shortly, developers can test the model in production without waiting for local weight downloads. The open‑source pledge eases lock‑in worries, giving teams cost control. GLM‑5.2 delivers immediate, high‑capacity assistance for today’s agentic workloads.