Download GLM-4.5

GLM-4.5 is an open-source large language model developed by Zhipu AI and Tsinghua University as part of the ChatGLM series. It delivers strong performance on a wide range of natural language tasks including reasoning, coding, mathematics, and multilingual understanding, with particular strength in Chinese-English bilingual contexts.

GLM-4.5 is designed to be run locally, making it a compelling choice for developers and researchers who want a capable AI model without relying on cloud APIs. It supports quantized variants that run efficiently on consumer-grade GPUs, lowering the hardware barrier for local LLM deployment.

The model supports long context windows and tool-calling capabilities, enabling agentic workflows and RAG (Retrieval-Augmented Generation) applications. It is available in multiple parameter sizes to suit different hardware configurations and use cases.

Key Features

• Strong bilingual performance in Chinese and English
• Excellent reasoning, coding, and mathematical problem-solving ability
• Long context window support for extended conversations and documents
• Tool-calling and function-calling support for agentic applications
• Available in multiple sizes with INT4/INT8 quantization for consumer GPUs
• Open-source weights released under a permissive license
• Compatible with llama.cpp, Ollama, and other local inference frameworks

How to Install

1. Click the download button to get the GLM-4.5 model weights or installer package.
2. Install a compatible inference framework such as Ollama (ollama pull glm4) or llama.cpp.
3. Load the model weights into your chosen inference engine following its documentation.
4. Run the model locally via the CLI or API endpoint exposed by your inference framework.
5. Optionally, integrate with a web UI like Open WebUI for a chat interface.