Can GLM 5 run on the Mac M3 Ultra 512GB?
Yes. At 8K context the best fit is Q3_K_M, using about 337 GB of the 384 GB usable on the Mac M3 Ultra 512GB. Estimated speed ceiling: ~40 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | 128K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|---|
| 2-bit (Q2_K) • | 258 GB Fits | 259 GB Fits | 261 GB Fits | 271 GB Fits | 53 |
| 3-bit (Q3_K_M) • | 337 GB Fits | 337 GB Fits | 340 GB Fits | 350 GB Tight | 40 |
| 4-bit (Q4_K_M) • | 426 GB No | 426 GB No | 429 GB No | 439 GB No | — |
| 5-bit (Q5_K_M) • | 500 GB No | 500 GB No | 502 GB No | 513 GB No | — |
| 6-bit (Q6_K) • | 578 GB No | 578 GB No | 581 GB No | 591 GB No | — |
| 8-bit (Q8_0) • | 747 GB No | 748 GB No | 750 GB No | 760 GB No | — |
| 16-bit (F16/BF16) • | 1406 GB No | 1406 GB No | 1408 GB No | 1419 GB No | — |
• real GGUF file size from unsloth/GLM-5-GGUF. Speed is a bandwidth ceiling (819 GB/s), not a benchmark. Methodology.
What fits instead
- No listed GPU with a list price runs GLM 5 fully in memory. The smallest listed hardware that does is the Mac M3 Ultra 512GB (512 GB).
- Largest popular model that fits the Mac M3 Ultra 512GB: Nex N2.6 (321B, MoE) at Q8_0, about 319 GB at 8K context.
"Runs" and "fits" mean the whole model in GPU memory at 8K context (4-bit or better for the model pick). Prices are US launch MSRPs, not today's street prices. "Popular" means the 60 most-downloaded models we track.
More
GLM 5 on every GPU and Mac · Every model that runs on the Mac M3 Ultra 512GB · Apple Silicon guide · Try other settings in the calculator
Guides: Best GPU for local LLMs by budget · Can I run gpt-oss, Llama 70B or Qwen3? · Mac vs PC for local LLMs
Data: Hugging Face Hub (zai-org/GLM-5), Apple specs. Model data last checked 2026-11-18; refreshed hourly.