GLM-6 — local AI model by Zhipu AI (Z.ai)
Written by Jakub Rusinowski · Last updated
GLM-6 has not been released. Z.ai's current generation is GLM-5.3 (August 2026), alongside GLM-5.3-Flash. The listings here were withdrawn; they redirect to the nearest real model.
Variants
The smallest GLM-6 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-6 9B 9B | ~6.2 GB | 128,000 | ollama run glm-6 | |
| GLM-6 355B-A32B 355B (32B active) | ~215.1 GB | 200,000 | ollama run glm-6 |
Memory is quantized weights plus overhead at Q4_K_M, from the same engine as the GPU & VRAM checker.
How to run GLM-6 locally
Install Ollama, then pull the tag.
ollama run glm-6Pick 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-6 9B, GLM-6 355B-A32BRecommended GPU
The cheapest catalogued GPU that runs GLM-6 locally (min 6 GB VRAM) is the Intel Arc B570 (10 GB).
Can I run GLM-6 on my GPU?
- GLM-6 on AMD Radeon RX 9060 XT 8GB
- GLM-6 on Apple M3 Ultra
- GLM-6 on Intel Arc B570
- GLM-6 on Intel Arc B580
- GLM-6 on NVIDIA GeForce RTX 3060 (12GB)
- GLM-6 on NVIDIA GeForce RTX 3070
- GLM-6 on NVIDIA GeForce RTX 3070 Ti
- GLM-6 on NVIDIA GeForce RTX 3080 (10GB)
- GLM-6 on NVIDIA GeForce RTX 4060
- GLM-6 on NVIDIA GeForce RTX 4070
- GLM-6 on NVIDIA GeForce RTX 4070 Super
- GLM-6 on NVIDIA GeForce RTX 4070 Ti
- GLM-6 on NVIDIA GeForce RTX 5060
- GLM-6 on NVIDIA GeForce RTX 5060 Ti 8GB
- GLM-6 on NVIDIA GeForce RTX 5070
GLM-6 — frequently asked questions
How much VRAM does GLM-6 need?
GLM-6 needs about 6 GB VRAM at Q4_K_M quantization for its smallest variant. Variants: GLM-6 9B (6 GB, Q4_K_M); GLM-6 355B-A32B (215 GB, Q4_K_M). On Apple Silicon, unified memory counts toward this requirement.
Can I run GLM-6 on an RTX 4090 (24 GB)?
Yes — GLM-6 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-6?
Q4_K_M is the best balance of quality and VRAM for GLM-6 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-6 with Ollama?
GLM-6 has no local Ollama tag — the published tag is cloud-hosted, so running it sends your prompts to a hosted GPU rather than your own machine.