GLM-5
GLM-5 publishes its weights, but at 744B parameters (40B active per token) no consumer machine holds the checkpoint. This page covers the access paths that work and the open models that actually run locally.
Why you can't run GLM-5 locally
At 744B parameters (40B active per token), a Q4-class build of GLM-5 would need roughly 446 GB — only a maxed-out 512GB Mac Studio could even hold the weights, with no real headroom. That figure is arithmetic, not opinion: 0.6 GB per billion parameters is the standard Q4 rule we apply across the whole catalog. The weights themselves are public (official source linked below); capacity, not licensing, is the wall.
What works instead: the hosted API. For most workloads, the open models below deliver the same job on hardware that fits under a desk.
Cloud/API only GLM-5, Zhipu AI's latest reasoning model with MoE architecture.
What to run locally instead of GLM-5
More Zhipu models
Frequently asked questions
Can I run GLM-5 locally?
Not on hardware you can buy. The weights are public, but GLM-5 is a 744B-parameter model (40B active), and a Q4-class build would need roughly 446 GB — only a maxed-out 512GB Mac Studio could even hold the weights, with no real headroom. Until the ecosystem ships a smaller official build, the API or a smaller open model is the practical path.
How do I access GLM-5?
Through the vendor's hosted API. The official source is linked on this page.
What is the best local alternative to GLM-5?
Qwen3 235B A22B is the strongest local model we track (235B, from 192 GB machines), with Qwen3.5 122B-A10B Instruct close behind. Both run on a single high-memory Mac or GPU — see their pages for exact hardware.
Cite this page
ModelFit: GLM-5 — specs, memory math and hardware verdicts. https://modelfit.io/models/glm-5/ (dataset updated 2026-09-03, CC BY 4.0).