GLM
A Chinese general model that began with autoregressive blank-infilling pretraining
WHAT IT IS
GLM is the general language-model series from Zhipu AI, its name standing for General Language Model and using an autoregressive blank-infilling pretraining objective. GLM-4 arrived in 2024, aimed at dialogue, function calling and agent scenarios and offering general generation in Chinese and other languages. The company grew out of Tsinghua University’s knowledge-engineering lab and was founded in 2019. The GLM family includes commercial versions plus some lightweight open ones, and is widely integrated into Chinese enterprise applications and agent frameworks.
Why it matters
It started from a blank-infilling pretraining objective different from the standard causal language model, and was early to build function calling and agent abilities into a Chinese model — one of China’s representatives of the open-plus-commercial route.
Key specs
- Architecture
- GLM (autoregressive blank infilling)
- Modality
- Text
- Open weights
- No (commercial version)
- Released
- 2023-03
Capabilities
Conversation & Instruction Following
Understand intent across turns and act on it
Text Generation
Continue a passage, one word at a time
Reasoning & Chain-of-Thought
Break a hard problem into intermediate steps
Tool Use & Function Calling
Let the model pick an API and fill its arguments
Related concepts
Transformer Architecture
Replacing word-by-word relay with a room where everyone speaks at once, so long-range dependencies are one hop away
Pretraining & Fine-tuning
Learn language first from vast unlabelled text, then specialise with little data — the most data-efficient paradigm in modern AI
Prompting & Alignment
Making a model helpful, honest and harmless is harder than simply making it bigger
Comparable products
ERNIE
2019A Chinese model that began with knowledge-enhanced pretraining, an early landmark version
Qwen
2023An open-weight family spanning many sizes, with multimodal versions
Baichuan
2023An open-weight general model aimed at Chinese-language use
Doubao
2023ByteDance’s general chat model and application
Kimi
2023A Chinese chat assistant known for long-context handling
Hunyuan
2023Tencent’s general model family, with open-weight versions
MiniMax-M
2025An open-weight reasoning model with hybrid attention and a million-token context
Step
2023A general model family aimed at multimodality and on-device use