Explore additional models
Frequently
asked questions
GLM-5.2 is Z.ai’s open-weight mixture-of-experts model, released in June 2026, with about 744B total parameters and roughly 40B active per token. It is built for agentic coding, multi-step tool use, and long-context reasoning, and ships under the MIT license. On Crusoe it is available for Serverless Fine-Tuning and Self-Serve Deployments in Crusoe Intelligence Foundry. You can also request a Tailored Deployment for bespoke inference optimization and a dedicated endpoint.
GLM-5.3 is the newer generation, with improved agentic coding and a 1,048,576-token context on Crusoe. GLM-5.2 is the earlier release, and the one to choose when you want to fine-tune a full-size GLM model on your own data. On Crusoe, GLM-5.3 is available for Serverless Inference and Self-Serve Deployments, while GLM-5.2 is available for Serverless Fine-Tuning and Self-Serve Deployments.
GLM-5.2 is billed by how you use it. Self-Serve Deployments are billed per GPU hour based on the GPU you choose, giving you predictable inference cost at production volume. Serverless Fine-Tuning is billed per 1M training tokens, where training tokens equal dataset tokens multiplied by epochs. Current rates are shown in the pricing module above.
GLM-5.2 is one of the strongest open-weight models for agentic coding, and Z.ai reports performance competitive with leading closed models on coding and agent benchmarks. Its MIT license permits commercial use and fine-tuning.
Yes, GLM-5.2 is available for Serverless Fine-Tuning with a 262,144-token supported training context. Upload a JSONL or Parquet dataset, run a LoRA job with pre-configured defaults, then deploy the result in one click to a dedicated endpoint through Self-Serve Deployments or download the weights under the MIT license. GLM-5.2 is not currently available for Serverless Inference.



%204.png)