Can You Run GLM-4.5 (355B MoE)? VRAM Requirements
GLM-4.5 (355B MoE) has 355B total parameters (MoE, 32B active per token). Running it at Q4_K_M (GGUF) with 16K context takes roughly 224 GB of memory (202 GB weights + 5.8 GB KV cache + runtime overhead). The minimum viable hardware is about Mac M3 Ultra 512GB. All numbers are planning estimates, verified 2026-07-27.
⚠ Some architecture details of this model are estimated — treat results as ballpark figures.
VRAM by quantization (16K context)
| Quantization | Weights | Total needed | Minimum hardware |
|---|---|---|---|
| FP16 / BF16 | 661 GB | 720 GB | 5× B200 192GB |
| FP8 | 331 GB | 363 GB | Mac M3 Ultra 512GB |
| Q8_0 (GGUF) | 350 GB | 384 GB | Mac M3 Ultra 512GB |
| Q6_K (GGUF) | 271 GB | 299 GB | Mac M3 Ultra 512GB |
| Q5_K_M (GGUF) | 235 GB | 259 GB | Mac M3 Ultra 512GB |
| MXFP4 | 175 GB | 195 GB | Mac M3 Ultra 512GB |
| Q4_K_M (GGUF)recommended | 202 GB | 224 GB | Mac M3 Ultra 512GB |
| Q3_K_M (GGUF) | 155 GB | 174 GB | AMD MI300X 192GB |
| Q2_K (GGUF) | 116 GB | 131 GB | AMD MI300X 192GB |
How context length changes memory (Q4_K_M (GGUF))
| Context | KV cache | Total needed |
|---|---|---|
| 4K | 1.4 GB | 219 GB |
| 16K | 5.8 GB | 224 GB |
| 64K | 23.0 GB | 241 GB |
| 128K | 46.0 GB | 264 GB |
KV cache grows linearly with context; FP8 KV cache halves it again (try it in the calculator below).
Customize the estimate
LLM VRAM Calculator — Can I Run It?
Estimate how much GPU memory (VRAM) you need to run open-weight LLMs like Kimi K3, DeepSeek R1, Qwen3, or Llama locally. Pick a model, quantization, and context length — the calculator adds up weights, KV cache, and runtime overhead, then shows which GPUs or Macs can fit it. Runs entirely in your browser: no upload, no signup.
355B total params · 32B active (MoE) · max 128K context · specs partially estimated
Recommended balance · ~0.61 bytes/param
Estimated memory needed
264 GB
Architecture details for this model are estimated — treat results as a ballpark.
Will it fit?
Assumes ~92% of device memory is usable. Multi-GPU counts are for tensor/pipeline parallel serving (vLLM, SGLang); Apple Silicon uses unified memory via llama.cpp or MLX.
FAQ
How much VRAM does GLM-4.5 (355B MoE) need?
About 224 GB at the recommended Q4_K_M (GGUF) quantization with 16K context: 202 GB for weights, 5.8 GB for KV cache, plus runtime overhead. Longer context grows the KV cache.
What is the minimum hardware for GLM-4.5 (355B MoE)?
Roughly Mac M3 Ultra 512GB, assuming 92% of device memory is usable. Lower-end setups can try more aggressive quantization (Q3/Q2) at a noticeable quality cost.
Can I run GLM-4.5 (355B MoE) on a Mac?
Yes. At Q4_K_M (GGUF) with 16K context it needs about 224 GB, which fits the unified memory of a Mac M3 Ultra 512GB via llama.cpp or MLX.
Which quantization should I use for GLM-4.5 (355B MoE)?
Q4_K_M is the recommended balance of size and quality; with headroom, Q6_K or Q8_0 reduce quality loss further.