A useful kind of curiosity.

What belongs here

GLM Handbook covers open language models and the engineering around them: model editions, hosted APIs, inference, context and evaluation. The name is a starting point for practical learning. The journal also covers relevant techniques across model families.

We favor questions with consequences: will the model fit, does the API support the workflow, what does the answer cost, and what evidence supports the claim?

The editorial desk

Articles are published under the GLM Handbook Editorial byline. This is a publication byline, not an invented individual biography. Guides are prepared with AI assistance and checked against the sources linked in each article. This site is not an official publication of Z.ai, Zhipu or the upstream model projects.

How to read the journal

Start with Models for terminology and memory, APIs for integration decisions, Inference for cost and latency, or Guides for practical workflows. Every article is available as ordinary HTML and plain Markdown.

Follow the RSS feed in your preferred reader. Publication dates reflect actual publication, and technical claims link to primary documentation where available.

Evidence and commercial interests

Our editorial and measurement policy explains source checks, supplied performance figures and the commissioning interest behind two API guides. We separate documentation checks, illustrative arithmetic and measurements rather than calling all three a benchmark.

Corrections

If a source changes, the affected article should be updated with a new source-check date and a correction note for material errors. Until then, use the check date alongside the current upstream documentation. We do not publish an unmonitored contact form or a newsletter signup that cannot deliver anything.