GLM

GLM is a family of language models created by the Chinese company Z.ai, formerly known as Zhipu AI. It does the same kind of work as the models behind ChatGPT or Claude (chatting, writing, analysing and coding), but with one important difference: many of its versions are published with open weights, so anyone can download them and run them on their own.

The name comes from General Language Model and it has been evolving for years: it started as ChatGLM and has kept releasing new generations, increasingly geared towards coding and working as an agent. What confuses beginners is that GLM can be used in three ways: in Z.ai's chat, through its API (the door that lets other programs use it) or by downloading the model. Being downloadable doesn't mean it fits on your computer: the big versions need very powerful machines. And, as with any cloud service, it's worth thinking about what data you give it.

GLM is one of the AIs I use besides Claude, which is my main tool. What its free version offers and where the limit is I explain in "Can you work professionally with free AI tools?". What I use it for, how it compares with Claude and OpenAI on my own tasks and what data I give it and what I don't, in "Kimi, GLM and MiniMax: the Chinese alternatives I use".