RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can GLM 4.5 Air run on the H200?

Yes. At 8K context the best fit is Q8_0, using about 112 GB of the 134 GB usable on the H200. Estimated speed ceiling: ~302 tokens/sec.

Memory by quant and context

Quant4K ctx8K ctx32K ctx128K ctxMax tok/s (est., 8K)
2-bit (Q2_K) •43.6 GB Fits44.3 GB Fits48.6 GB Fits65.9 GB Fits782
3-bit (Q3_K_M) •54.7 GB Fits55.4 GB Fits59.7 GB Fits77.0 GB Fits619
4-bit (Q4_K_M) •69.4 GB Fits70.1 GB Fits74.4 GB Fits91.7 GB Fits486
5-bit (Q5_K_M) •79.2 GB Fits79.9 GB Fits84.2 GB Fits101 GB Fits424
6-bit (Q6_K) •93.6 GB Fits94.3 GB Fits98.6 GB Fits116 GB Fits358
8-bit (Q8_0) •111 GB Fits112 GB Fits116 GB Fits133 GB Tight302
16-bit (F16/BF16) •207 GB No208 GB No212 GB No230 GB No—

• real GGUF file size from unsloth/GLM-4.5-Air-GGUF. Speed is a bandwidth ceiling (4800 GB/s), not a benchmark. Methodology.

More

GLM 4.5 Air on every GPU and Mac · Every model that runs on the H200 · Other models that fit 64 GB VRAM · Try other settings in the calculator

Data: Hugging Face Hub (zai-org/GLM-4.5-Air), NVIDIA specs. Data as of 2026-11-17.