NVIDIA NIM Enables Access to GLM-5.2 Model

Access to a 753 billion parameter model with a 1,000,000-token context window removes a major infrastructure barrier for long-horizon research and tool-enabled agents. Reporting: GLM-5.2 is an open-source flagship model from Z.ai with 753B parameters and a 1,000,000-token context window, per NVIDIA's API documentation (z-ai / glm-5.2) and the Build.NVIDIA model listing. NVIDIA's NIM model catalog lists GLM-5.2 with a free endpoint and downloadable options on Build.NVIDIA (models page). PlainEnglish coverage also highlights free API access via NVIDIA NIM, characterising the endpoint as free and developer-facing. NVIDIA's docs additionally describe GLM-5.2 as a mixture-of-experts model using IndexShare sparse attention and list the model as commercially usable, with release traces on Build.NVIDIA and Hugging Face.