GLM-4.5-Air 106B: how much VRAM?
In Q4_K_M with 8k tokens of context, GLM-4.5-Air 106B needs about 69 GB of memory. The cheapest option: Mini-PC Ryzen AI Max+ 395 128 GB (€2,300 for the complete machine).
Model sheet
vllm serve zai-org/GLM-4.5-AirWeights on Hugging Face: zai-org/GLM-4.5-Air
VRAM by precision and context
| Precision | 4k | 8k | 32k | 128k |
|---|---|---|---|---|
| FP16 | 224 GB | 224 GB | 229 GB | 246 GB |
| FP8 | 112 GB | 113 GB | 117 GB | 135 GB |
| Q8_0 | 119 GB | 120 GB | 124 GB | 141 GB |
| Q6_K | 92 GB | 93 GB | 97 GB | 115 GB |
| Q4_K_M | 68 GB | 69 GB | 73 GB | 90 GB |
| Q3_K_M | 56 GB | 56 GB | 61 GB | 78 GB |
One user. Each simultaneous user adds their own context cache.
Frequently asked questions
How much VRAM does GLM-4.5-Air 106B need?
With an 8k-token context and one user, GLM-4.5-Air 106B needs about 69 GB in Q4, 120 GB in Q8 and 224 GB in FP16.
What is the cheapest hardware to run GLM-4.5-Air 106B?
In Q4, the cheapest option in our catalogue to run GLM-4.5-Air 106B is Mini-PC Ryzen AI Max+ 395 128 GB, about €2,300 for the complete machine.
How fast does GLM-4.5-Air 106B run?
Expect around 19 tokens per second for one user on Mini-PC Ryzen AI Max+ 395 128 GB in Q4 (estimate based on memory bandwidth).
On which hardware?
| Hardware | Cards | Speed | Machine |
|---|---|---|---|
| Mini-PC Ryzen AI Max+ 395 128 GB | 1 is enough | ~19 tok/s | €2,300 |
| NVIDIA DGX Spark 128 GB | 1 is enough | ~20 tok/s | €4,000 |
| Mac Studio M4 Max 128 GB | 1 is enough | ~41 tok/s | €4,400 |
| RTX 3090 24 GB (used) | 3 cards | ~158 tok/s | €5,050 |
| Radeon RX 7900 XTX 24 GB | 3 cards | ~162 tok/s | €5,650 |
| RTX 4090 24 GB | 3 cards | ~170 tok/s | €8,800 |
| RTX PRO 6000 Blackwell 96 GB | 1 is enough | ~134 tok/s | €9,900 |
| RTX 5090 32 GB | 3 cards | ~302 tok/s | €10,600 |
| Mac Studio M3 Ultra 512 GB | 1 is enough | ~61 tok/s | €11,500 |
| RTX 6000 Ada 48 GB | 2 cards | ~108 tok/s | €14,900 |
| A100 80 GB | 1 is enough | ~153 tok/s | €17,000 |
| L40S 48 GB | 2 cards | ~97 tok/s | €21,000 |
| H100 80 GB | 1 is enough | ~251 tok/s | €34,000 |
| H200 141 GB | 1 is enough | ~359 tok/s | €39,000 |
| B200 192 GB | 1 is enough | ~599 tok/s | €48,000 |
| 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 4070 Ti SUPER 16 GB | doesn’t fit | — | — |
| RTX 4080 SUPER 16 GB | doesn’t fit | — | — |
| RTX 5080 16 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: Mini-PC Ryzen AI Max+ 395 128 GB · €2,300
From: Mini-PC Ryzen AI Max+ 395 128 GB · €2,300
From: Mac Studio M3 Ultra 512 GB · €11,500