Z.ai lists GLM-5.3 for deployment across open inference frameworks

Z.ai has published GLM-5.3 as an open-weight model, according to its Hugging Face listing. The company says the release uses the same base model as GLM-5.2 and attributes its gains to post-training. It lists deployment support for SGLang, vLLM, TokenSpeed, Transformers, KTransformers and Unsloth, plus several Ascend NPU inference frameworks. Z.ai also says users can set a three-level reasoning-effort parameter, which defaults to its maximum setting.