Hardware guide · October 2026
What GPU do you need to run GLM-5.3?
GLM-5.3 is a 744B (40B active) mixture-of-experts model. Here is exactly how much memory it needs and which GPUs and Macs run it well.
The short answer
To run GLM-5.3 at Q4_K_M with 8K tokens of context you need about 466GB of VRAM (451.0GB of weights plus KV cache and overhead). On a Mac or other unified-memory system, the M3 Ultra 512GB runs it at ~13 tokens/sec.
- VRAM at Q4_K_M
- ~466GB
- GPUs that run it
- 2
- Parameters
- 744B
- CPU only (est.)
- ~1.8 tok/s
Our recommendations
The best GPUs for GLM-5.3
Mac / unified memory
M3 Ultra 512GB512GB unified · $9,499 MSRP
- Quant
- Q3_K_M
- Speed (est.)
- ~13 tok/s
Lots of memory per dollar and silent, but slower than a GPU with the same capacity.
We may earn a commission when you buy through links on this page. It never changes what we recommend. How we make money
Memory
GLM-5.3 VRAM requirements
Memory needed at each quantization level. Q4_K_M is the best balance of size and quality for most people; move up to Q5 or Q6 if you have the room.
| Quantization | Quality | Weights | Total @ 8K ctx | Total @ 32K ctx |
|---|---|---|---|---|
| Q2_K | Rough | 279 GB | 289 GB | 291 GB |
| Q3_K_M | Usable | 363 GB | 375 GB | 377 GB |
| Q4_K_M | RecommendedDefault | 451 GB | 466 GB | 468 GB |
| Q5_K_M | Very good | 530 GB | 547 GB | 550 GB |
| Q6_K | Near-lossless | 614 GB | 634 GB | 636 GB |
| Q8_0 | Practically lossless | 791 GB | 816 GB | 818 GB |
| BF16 | Full precision | 1,488 GB | 1,534 GB | 1,536 GB |
Context
How context length changes the memory you need
Every token in the context window is stored in the KV cache. At Q4_K_M, with a 16-bit KV cache, GLM-5.3 needs:
| Context | KV cache | Total memory | With 8-bit KV |
|---|---|---|---|
| 4K tokens | 0.4 GB | 465.5 GB | 465.4 GB |
| 8K tokens | 0.7 GB | 465.9 GB | 465.5 GB |
| 16K tokens | 1.5 GB | 466.7 GB | 465.9 GB |
| 32K tokens | 2.9 GB | 468.1 GB | 466.7 GB |
| 64K tokens | 5.9 GB | 471.1 GB | 468.1 GB |
| 128K tokens | 11.8 GB | 477.0 GB | 471.1 GB |
| 256K tokens | 23.6 GB | 488.7 GB | 477.0 GB |
Multi-head latent attention with DeepSeek-style sparse attention. Assuming the GLM-5 layout of 78 layers, each caching a 576-value latent (512 compressed plus 64 RoPE): 78 x 576 x 2 bytes = 89,856 bytes per token, about 94GB at 1M tokens, plus a small sparse-attention indexer cache. Layer count not confirmed from the config.
Compatibility
Every GPU that runs GLM-5.3
Each GPU at the highest-quality quantization that fits with 8K context, fastest first. Speeds are single-user estimates.
| GPU | Memory | Quant | Speed | Fit | Price |
|---|---|---|---|---|---|
| M5 Ultra 512GB | 512GB unified | Q3_K_M | ~19 tok/s | Fits, tight | — |
| M3 Ultra 512GB | 512GB unified | Q3_K_M | ~13 tok/s | Fits, tight | $9,499 |
Tips
Getting the most out of GLM-5.3
- Start at Q4_K_M. It keeps nearly all of the quality at about a third of the 16-bit size. Step up to Q5_K_M or Q6_K only if it still fits with the context you need.
- Budget memory for context. The KV cache grows with every token. Quantizing it to 8-bit roughly halves that cost with little quality loss.
- Offload experts, not layers. For mixture-of-experts models, llama.cpp's
--n-cpu-moe(or-ottensor overrides) keeps attention on the GPU and moves expert weights to system RAM. That keeps speed usable when the model does not fit in VRAM. - Pick the right runtime. Ollama and LM Studio are easiest. llama.cpp gives the most control. vLLM or SGLang are best when several people or agents share one machine.
Want to try different quantizations, context lengths, or GPU counts? Use our VRAM calculator, or read the GLM-5.3 overview for setup commands and benchmarks.
Frequently asked questions
How much VRAM do I need to run GLM-5.3?
About 466GB at Q4_K_M with an 8K context window. At Q8_0 (near-lossless) it needs about 816GB, and at Q3_K_M about 375GB with some quality loss. Longer contexts add 0.09GB per 1K tokens with a 16-bit KV cache.
Can I run GLM-5.3 on a 24GB GPU?
No. GLM-5.3 needs about 466GB at Q4_K_M, which does not fit on a 24GB card like the RTX 3090 or RTX 4090, even with CPU offload on a 64GB system.
Can I run GLM-5.3 on a Mac?
Yes. With 512GB of unified memory or more (for example the M3 Ultra 512GB), macOS can give the GPU enough memory to run it at Q3_K_M, at about ~13 tokens/sec. MLX builds are usually fastest on Apple silicon.
Can I run GLM-5.3 on CPU only, without a GPU?
Technically yes on a server with enough RAM, but at about 466GB it needs a multi-channel server platform and will be slow on a desktop.