GLM-4.5 355B: how much VRAM?
In Q4_K_M with 8k tokens of context, GLM-4.5 355B needs about 226 GB of memory. The cheapest option: Mac Studio M3 Ultra 512 GB (€11,500 for the complete machine).
Model sheet
vllm serve zai-org/GLM-4.5Weights on Hugging Face: zai-org/GLM-4.5
VRAM by precision and context
| Precision | 4k | 8k | 32k | 128k |
|---|---|---|---|---|
| FP16 | 747 GB | 748 GB | 756 GB | 784 GB |
| FP8 | 374 GB | 376 GB | 383 GB | 412 GB |
| Q8_0 | 397 GB | 398 GB | 405 GB | 434 GB |
| Q6_K | 307 GB | 308 GB | 316 GB | 344 GB |
| Q4_K_M | 225 GB | 226 GB | 234 GB | 262 GB |
| Q3_K_M | 184 GB | 185 GB | 193 GB | 221 GB |
One user. Each simultaneous user adds their own context cache.
Frequently asked questions
How much VRAM does GLM-4.5 355B need?
With an 8k-token context and one user, GLM-4.5 355B needs about 226 GB in Q4, 398 GB in Q8 and 748 GB in FP16.
What is the cheapest hardware to run GLM-4.5 355B?
In Q4, the cheapest option in our catalogue to run GLM-4.5 355B is Mac Studio M3 Ultra 512 GB, about €11,500 for the complete machine.
How fast does GLM-4.5 355B run?
Expect around 23 tokens per second for one user on Mac Studio M3 Ultra 512 GB in Q4 (estimate based on memory bandwidth).
On which hardware?
| Hardware | Cards | Speed | Machine |
|---|---|---|---|
| Mac Studio M3 Ultra 512 GB | 1 is enough | ~23 tok/s | €11,500 |
| RTX PRO 6000 Blackwell 96 GB | 3 cards | ~115 tok/s | €29,800 |
| A100 80 GB | 3 cards | ~130 tok/s | €39,000 |
| L40S 48 GB | 5 cards | ~92 tok/s | €43,500 |
| H200 141 GB | 2 cards | ~205 tok/s | €72,000 |
| H100 80 GB | 3 cards | ~214 tok/s | €90,000 |
| B200 192 GB | 2 cards | ~341 tok/s | €90,000 |
| RTX 3060 12 GB | doesn’t fit | — | — |
| RTX 4060 Ti 16 GB | doesn’t fit | — | — |
| Radeon RX 7800 XT 16 GB | doesn’t fit | — | — |
| RTX 4070 12 GB | doesn’t fit | — | — |
| RTX 3090 24 GB (used) | doesn’t fit | — | — |
| RTX 4070 Ti SUPER 16 GB | doesn’t fit | — | — |
| Radeon RX 7900 XTX 24 GB | doesn’t fit | — | — |
| RTX 4080 SUPER 16 GB | doesn’t fit | — | — |
| RTX 5080 16 GB | doesn’t fit | — | — |
| RTX 4090 24 GB | doesn’t fit | — | — |
| Mini-PC Ryzen AI Max+ 395 128 GB | doesn’t fit | — | — |
| RTX 5090 32 GB | doesn’t fit | — | — |
| NVIDIA DGX Spark 128 GB | doesn’t fit | — | — |
| Mac Studio M4 Max 128 GB | doesn’t fit | — | — |
| RTX 6000 Ada 48 GB | doesn’t fit | — | — |
Estimated complete machine: cards + host PC (€900 for 1–2 consumer cards, €2,800 beyond, a €6,000 server for datacenter cards). Macs and mini-PCs are sold complete.
Similar models
From: Mini-PC Ryzen AI Max+ 395 128 GB · €2,300
From: Mac Studio M3 Ultra 512 GB · €11,500
From: Mac Studio M3 Ultra 512 GB · €11,500