GLM-Z1 32B (Reasoning) — VRAM, speed & local setup
Written by Jakub Rusinowski · Last updated
Reasoning-specialized GLM variant. Near-GPT-5.2 on the AI Index. Best for math, logic, and scientific reasoning tasks.
GLM-Z1 32B (Reasoning) needs about 20 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache. On Apple Silicon that figure comes out of unified memory.
VRAM and speed by quantization
Quoted against NVIDIA RTX 4090 (24 GB). Weights plus overhead only: this model's architecture is unpublished, so no KV cache is added.
| Quant | Memory | VRAM | Speed (est.) | Fit |
|---|---|---|---|---|
| Q2_K 2.63 bpw | 11.3 GB | ~58 tok/s | Fits | |
| Q3_K_M 3.41 bpw | 14.4 GB | ~46 tok/s | Fits | |
| Q4_K_M 4.83 bpw | 20.1 GB | ~34 tok/s | Fits | |
| Q5_K_M 5.67 bpw | 23.5 GB | ~30 tok/s | Tight | |
| Q6_K 6.56 bpw | 27 GB | ~4 tok/s | Offload | |
| Q8_0 8.50 bpw | 34.8 GB | ~3 tok/s | Offload | |
| F16 16.00 bpw | 64.8 GB | — | Too big |
Black marker = usable memory on the NVIDIA RTX 4090 (24 GB). Estimates from the memory-bandwidth roofline on the methodology page. GLM-Z1 32B (Reasoning) VRAM calculator →
Get GLM-Z1 32B (Reasoning) running
The cheapest catalogued GPU that runs GLM-Z1 32B (Reasoning) is the AMD Radeon RX 7900 XTX (24 GB).
or compare on Vast.ai from $0.35/hr (typical low · varies)
As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.
How to run GLM-Z1 32B (Reasoning)
Install Ollama, then run:
ollama run glm-z1:32bSpecifications
Preview — The model is released, but these specs are thin or rest on a single source. Individual fields may be wrong.
Preview — The model is released, but these specs are thin or rest on a single source. Individual fields may be wrong.
- Parameters
- 32 Billion
- Context window
- 128,000
- Architecture
- Dense
- Provider
- Zhipu AI (Z.ai)
- Licence
- Apache 2.0
- Practical deployment
- Single consumer GPU
- Specified at
- Q4_K_M
- System RAM
- 48 GB
- Record updated
- 2026-02-14
Commercial use permitted. No usage restrictions beyond attribution.
Quality and use cases
Scores as published by the model’s authors or an independent evaluator — quality, not throughput, and not measured by us.
Can I run GLM-Z1 32B (Reasoning) on my GPU?
- GLM-Z1 32B (Reasoning) on AMD Radeon RX 7900 XT
- GLM-Z1 32B (Reasoning) on AMD Radeon RX 7900 XTX
- GLM-Z1 32B (Reasoning) on AMD Radeon RX 9060 XT 8GB
- GLM-Z1 32B (Reasoning) on Apple M1 Pro
- GLM-Z1 32B (Reasoning) on Apple M2 Pro
- GLM-Z1 32B (Reasoning) on Apple M3 Pro
- GLM-Z1 32B (Reasoning) on Apple M4
- GLM-Z1 32B (Reasoning) on Apple M5
- GLM-Z1 32B (Reasoning) on Intel Arc B570
- GLM-Z1 32B (Reasoning) on Intel Arc B580
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 3060 (12GB)
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 3070
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 3070 Ti
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 3080 (10GB)
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 3090
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 4060
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 4070
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 4070 Super
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 4070 Ti
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 4090
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 5060
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 5060 Ti 8GB
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 5070
- GLM-Z1 32B (Reasoning) on NVIDIA GeForce RTX 5090
- GLM-Z1 32B (Reasoning) on NVIDIA L40S
- GLM-Z1 32B (Reasoning) on NVIDIA RTX 6000 Ada Generation
Other GLM-4.7 / GLM-Z1 sizes
GLM-Z1 32B (Reasoning) — frequently asked questions
How much VRAM does GLM-Z1 32B (Reasoning) need?
About 20 GB at Q4_K_M — quantized weights plus framework overhead, before any KV cache. The cache grows with context length and is added on top; the table above folds it in. Apple Silicon counts unified memory toward the same figure.
Does GLM-Z1 32B (Reasoning) run on an RTX 4090 (24 GB)?
Yes. GLM-Z1 32B (Reasoning) needs about 20 GB at Q4_K_M, inside a 24 GB card, at an estimated 34 tokens/sec.
How do I run GLM-Z1 32B (Reasoning) locally?
Install Ollama and run `ollama run glm-z1:32b`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.
What other sizes does GLM-4.7 / GLM-Z1 come in?
GLM-4 9B (6 GB), GLM-Z1 32B (Reasoning) (20 GB), GLM-4.7-Flash 30B-A3B (19 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.