Can a NVIDIA H200 (141GB) run GLM-4.5 Air (MoE)?

Yes — GLM-4.5 Air (MoE) runs at Q6_K (99.0 GB) at about 167–251 tok/s on a NVIDIA H200 (141GB).

How much memory does GLM-4.5 Air (MoE) need?

GLM-4.5 Air (MoE) needs about 46.9 GB at 4K context at its smallest quantization. A NVIDIA H200 (141GB) has 141.0 GB.

Yes — GLM-4.5 Air (MoE) runs on a NVIDIA H200 (141GB) at Q6_K

Yes — it runs great.

Runs great

Best fit: GLM-4.5 Air (MoE) at Q6_K · 99.0 GB weights · 167–251 tok/s on a NVIDIA H200 (141GB).

Arena Text Overall · style-controlled

1,373Q6_K untested

glm-4.5-air · published configuration reference · Q6_K not tested

Arena Text Overall (style-controlled), exact published entry glm-4.5-air, 2026-10-08: reference rating 1,373.1. Published confidence bounds 1,368.9–1,377.4. 30,658 votes. Published glm-4.5-air configuration reference. Reasoning mode is not disclosed by the Arena dataset. Tested precision, runtime and sampling settings are not disclosed in this dataset. Selected Q6_K has not been measured by this Arena entry. No quantization quality adjustment is applied.

Snapshot published 2026-10-08 · snapshot imported 2026-10-11 UTC · validated last-good data · manual refresh

Arena leaderboard dataset · Arena / lmarena-ai, Arena Leaderboard Dataset (CC BY 4.0). Selected Text Overall style-controlled rows; field names adapted, ratings unchanged. · dataset CC BY 4.0 · revision e16f90232986

Total memory at 4K context: 100.6 GB; Weights + KV cache + runtime overhead
Estimated generation: 167–251 tok/s
Estimated first token: ~0.6s

Every quantization on this machine

On small screens, each quant lists every metric below its name. On wider screens, scroll sideways for total memory, speed and fit. Keyboard: focus the table, then use the arrow keys. Quant names stay visible.

GLM-4.5 Air (MoE) quantizations on NVIDIA H200 (141GB), estimated at 4K context. Weights and total memory are shown separately.
Quant	Weights	Total memory	Speed	Fit
Q2_K	45.3 GB	46.9 GB	231–346 tok/s	Runs great
Q3_K_M	57.2 GB	58.8 GB	213–319 tok/s	Runs great
Q4_K_M	73.0 GB	74.6 GB	193–289 tok/s	Runs great
Q5_K_M	83.5 GB	85.1 GB	182–272 tok/s	Runs great
Q6_K	99.0 GB	100.6 GB	167–251 tok/s	Runs great
Q8_0	117.5 GB	119.0 GB	153–229 tok/s	Tight / slow
FP16	221.0 GB	222.6 GB	— not resident	Won't fit

Open the full calculator →Everything a NVIDIA H200 (141GB) runs GLM-4.5 Air (MoE) on other machines What to buy to run it →

Ran GLM-4.5 Air (MoE) on a NVIDIA H200 (141GB)? Add your real numbers →