Can a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) run Qwen3 30B-A3B (MoE)?

Yes — Qwen3 30B-A3B (MoE) runs at FP16 (61.1 GB) at about 222–334 tok/s on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB).

How much memory does Qwen3 30B-A3B (MoE) need?

Qwen3 30B-A3B (MoE) needs about 12.5 GB at 4K context at its smallest quantization. A NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) has 144.0 GB.

Yes — Qwen3 30B-A3B (MoE) runs on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) at FP16

Yes — it runs great.

Runs great

Best fit: Qwen3 30B-A3B (MoE) at FP16 · 61.1 GB weights · 222–334 tok/s on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB).

Arena Text Overall · style-controlled

1,382FP16 untested

qwen3-30b-a3b-instruct-2507 · published configuration reference · FP16 not tested

Arena Text Overall (style-controlled), exact published entry qwen3-30b-a3b-instruct-2507, 2026-10-08: reference rating 1,382.4. Published confidence bounds 1,377.5–1,387.3. 23,325 votes. Published qwen3-30b-a3b-instruct-2507 configuration reference. Reasoning mode is not disclosed by the Arena dataset. Tested precision, runtime and sampling settings are not disclosed in this dataset. Selected FP16 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: 62.3 GB; Weights + KV cache + runtime overhead
Estimated generation: 222–334 tok/s
Estimated first token: ~0.2s

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.

Qwen3 30B-A3B (MoE) quantizations on NVIDIA GH200 Grace Hopper (144GB HBM + 480GB), estimated at 4K context. Weights and total memory are shown separately.
Quant	Weights	Total memory	Speed	Fit
Q2_K	11.3 GB	12.5 GB	318–477 tok/s	Runs great
Q3_K_M	14.7 GB	15.9 GB	309–464 tok/s	Runs great
Q4_K_M	18.6 GB	19.8 GB	299–449 tok/s	Runs great
Q5_K_M	21.7 GB	22.9 GB	292–438 tok/s	Runs great
Q6_K	25.1 GB	26.3 GB	284–426 tok/s	Runs great
Q8_0	32.5 GB	33.7 GB	269–403 tok/s	Runs great
FP16	61.1 GB	62.3 GB	222–334 tok/s	Runs great

Open the full calculator →Everything a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) runs Qwen3 30B-A3B (MoE) on other machines What to buy to run it →

Ran Qwen3 30B-A3B (MoE) on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB)? Add your real numbers →