GLM-5.2 Quantized Model Hits 16 tok/s on M3 Ultra 512GB
antirez · x · 2026-07-06
A user ran the GLM-5.2 q4 quantized model (ds4-eval version) on an Apple M3 Ultra 512GB device, achieving an inference speed of roughly 16 token/s. The developer plans to update older branches to introduce batch inference for further performance gains. This serves as a real-world benchmark for deploying Zhipu's GLM series models locally on high-end Apple silicon.
Related event: Quantized GLM-5.2 Runs Locally on M3 Ultra at 16 Tokens/s(2 posts)→
More from Infra
- WSJ: Nvidia is in talks to backstop about $250 billion of OpenAI's data center plan — KateClarkTweets · 2026-07-27
- YC talk on BCI x AI says infrastructure is what really determines speed — garrytan · 2026-07-27
- A 13B model ran on a no-GPU PC by paging weights from SSD via llama.cpp — ID_R_McGregor · 2026-07-27
- llama.cpp warns that GGUFs made before a recent change must be regenerated — EconomySerious · 2026-07-27
- RTX 5090 local tests show Qwen Q6 can drop to 15 tok/s at 80k context — LFAdvice7984 · 2026-07-27
- Surprising Ubuntu Setup: NVIDIA 5090 PC Becomes the Easiest AI Rig — _xjdr · 2026-07-27