Can a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) run Gemma 4 26B-A4B (MoE)?

Yes — Gemma 4 26B-A4B (MoE) runs at FP16 (50.5 GB) at about 199–299 tok/s on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB).

How much memory does Gemma 4 26B-A4B (MoE) need?

Gemma 4 26B-A4B (MoE) needs about 11.6 GB at 4K context at its smallest quantization. A NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) has 144.0 GB.

Yes — Gemma 4 26B-A4B (MoE) runs on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) at FP16

Yes — it runs great.

Runs great

Best fit: Gemma 4 26B-A4B (MoE) at FP16 · 50.5 GB weights · 199–299 tok/s on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB).

Arena Text Overall · style-controlled

1,438FP16 untested

gemma-4-26b-a4b · published configuration reference · FP16 not tested

Arena Text Overall (style-controlled), exact published entry gemma-4-26b-a4b, 2026-10-08: reference rating 1,437.6. Published confidence bounds 1,430.2–1,445.1. 6,092 votes. Published gemma-4-26b-a4b 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: 51.6 GB; Weights + KV cache + runtime overhead
Estimated generation: 199–299 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.

Gemma 4 26B-A4B (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	10.6 GB	11.6 GB	306–459 tok/s	Runs great
Q3_K_M	12.7 GB	13.8 GB	297–446 tok/s	Runs great
Q4_K_M	17.0 GB	18.0 GB	282–423 tok/s	Runs great
Q5_K_M	21.2 GB	22.2 GB	268–402 tok/s	Runs great
Q6_K	23.2 GB	24.3 GB	262–393 tok/s	Runs great
Q8_0	26.9 GB	28.0 GB	251–377 tok/s	Runs great
FP16	50.5 GB	51.6 GB	199–299 tok/s	Runs great

Open the full calculator →Everything a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) runs Gemma 4 26B-A4B (MoE) on other machines What to buy to run it →

Ran Gemma 4 26B-A4B (MoE) on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB)? Add your real numbers →