Z.aiGLM-5
GLM-5.2
A long-context GLM release with a shared sparse-attention indexer.
Practical use
Practical guidance. For repository work, compare the model with the same files and task history used by newer releases. A large context is most useful when the input includes the decisions and constraints needed to judge the final edit.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API, Downloadable |
| Model identifiers | glm-5.2zai-org/GLM-5.2 |
| Architecture | IndexShare reuses an indexer across four sparse-attention layers. The release also updates multi-token prediction.GLM-5.2 model card |
| API limits | Text input and output; one-million-token context and 128,000 output tokens.GLM-5.2 guide |
| Licence | MIT.GLM-5.2 model card |
Related models
A text model with stronger post-training on the GLM-5.2 base.
Launch date not confirmedAPI · DownloadableA sparse text model for reasoning and software engineering.
Launch date not confirmedAPI · DownloadableA GLM release trained for sustained multi-step work.
Launch date not confirmedAPI · Downloadable