GLM-5 / GLM-5.1 — local AI model by Zhipu AI (Z.ai)
Written by Jakub Rusinowski · Last updated
Zhipu AI's (Z.ai) fifth-generation model series, released under MIT license for maximum flexibility. GLM-5 and GLM-5.1 excel at agentic workflows, tool calling, and long-context reasoning. The MIT license makes them one of the most commercially permissive frontier-class models available, with zero restrictions on commercial use or distribution. Both ship only as a 744B Mixture-of-Experts model (40B active, SWE-bench Verified leader) trained entirely on Huawei Ascend hardware: a data-center model, with no smaller consumer checkpoint. The 9B, 32B and 72B sizes listed here earlier were never released.
Variants
The smallest GLM-5 / GLM-5.1 variant needs about 6 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache.
| Model | VRAM at Q4 | VRAM | Context | Run it |
|---|---|---|---|---|
| GLM-5 9B 9B | ~6.2 GB | 128,000 | ollama run hf.co/THUDM/GLM-5-9B-Chat-Q4_K_M | |
| GLM-5 32B 32B | ~20.1 GB | 128,000 | ollama run hf.co/THUDM/GLM-5-32B-Chat-Q4_K_M | |
| GLM-5.1 72B 72B | ~44.3 GB | 128,000 | ollama run hf.co/THUDM/GLM-5.1-72B-Chat-Q4_K_M | |
| GLM-5 744B → 744B | ~450 GB | 200,000 | ollama run glm-5 |
Memory is quantized weights plus overhead at Q4_K_M, from the same engine as the GPU & VRAM checker.
How to run GLM-5 / GLM-5.1 locally
Install Ollama, then pull the tag.
ollama run hf.co/THUDM/GLM-5-9B-Chat-Q4_K_MPick a size above for its own VRAM figure, speed estimate and install command.
Licence
Commercial use permitted. No usage restrictions beyond attribution.
Applies to: GLM-5 9B, GLM-5 32B, GLM-5.1 72B, GLM-5 744BRecommended GPU
The cheapest catalogued GPU that runs GLM-5 / GLM-5.1 locally (min 6 GB VRAM) is the Intel Arc B570 (10 GB).
Can I run GLM-5 / GLM-5.1 on my GPU?
- GLM-5 / GLM-5.1 on AMD Radeon RX 7900 XTX
- GLM-5 / GLM-5.1 on AMD Radeon RX 9060 XT 8GB
- GLM-5 / GLM-5.1 on AMD Ryzen AI Max+ 395
- GLM-5 / GLM-5.1 on Apple M1 Max
- GLM-5 / GLM-5.1 on Apple M1 Pro
- GLM-5 / GLM-5.1 on Apple M2 Max
- GLM-5 / GLM-5.1 on Apple M2 Pro
- GLM-5 / GLM-5.1 on Apple M3 Max
- GLM-5 / GLM-5.1 on Apple M3 Pro
- GLM-5 / GLM-5.1 on Apple M4
- GLM-5 / GLM-5.1 on Apple M5
- GLM-5 / GLM-5.1 on Apple M5 Pro
- GLM-5 / GLM-5.1 on Intel Arc B570
- GLM-5 / GLM-5.1 on Intel Arc B580
- GLM-5 / GLM-5.1 on NVIDIA A100 80GB (PCIe)
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 3060 (12GB)
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 3070
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 3070 Ti
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 3080 (10GB)
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 3090
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 4060
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 4070
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 4070 Super
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 4070 Ti
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 4090
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 5060
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 5060 Ti 8GB
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 5070
- GLM-5 / GLM-5.1 on NVIDIA GeForce RTX 5090
- GLM-5 / GLM-5.1 on NVIDIA H100 80GB (PCIe)
- GLM-5 / GLM-5.1 on NVIDIA L40S
- GLM-5 / GLM-5.1 on NVIDIA RTX 6000 Ada Generation
- GLM-5 / GLM-5.1 on NVIDIA RTX PRO 6000 Blackwell
GLM-5 / GLM-5.1 — frequently asked questions
How much VRAM does GLM-5 / GLM-5.1 need?
GLM-5 / GLM-5.1 needs about 6 GB VRAM at Q4_K_M quantization for its smallest variant. Variants: GLM-5 9B (6 GB, Q4_K_M); GLM-5 32B (20 GB, Q4_K_M); GLM-5.1 72B (44 GB, Q4_K_M); GLM-5 744B (450 GB, Q4_K_M). On Apple Silicon, unified memory counts toward this requirement.
Can I run GLM-5 / GLM-5.1 on an RTX 4090 (24 GB)?
Yes — GLM-5 / GLM-5.1 runs on an RTX 4090 (24 GB) and other 24 GB cards such as the RTX 3090. Smaller variants also fit comfortably on 8–16 GB GPUs at Q4_K_M.
What quantization should I use for GLM-5 / GLM-5.1?
Q4_K_M is the best balance of quality and VRAM for GLM-5 / GLM-5.1 in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
How do I run GLM-5 / GLM-5.1 with Ollama?
Install Ollama, then run: ollama run hf.co/THUDM/GLM-5-9B-Chat-Q4_K_M. This downloads GLM-5 / GLM-5.1 and starts a local, OpenAI-compatible endpoint — no internet connection is needed after the initial download.