Can a NVIDIA L40S (48GB) run Gemma 4 31B?

Yes — Gemma 4 31B runs at Q8_0 (32.6 GB) at about 18–27 tok/s on a NVIDIA L40S (48GB).

How much memory does Gemma 4 31B need?

Gemma 4 31B needs about 13.7 GB at 4K context at its smallest quantization. A NVIDIA L40S (48GB) has 48.0 GB.

Yes — Gemma 4 31B runs on a NVIDIA L40S (48GB) at Q8_0

Yes — it runs great.

Runs great

Best fit: Gemma 4 31B at Q8_0 · 32.6 GB weights · 18–27 tok/s on a NVIDIA L40S (48GB).

Arena Text Overall · style-controlled

1,452Q8_0 untested

gemma-4-31b · published configuration reference · Q8_0 not tested

Arena Text Overall (style-controlled), exact published entry gemma-4-31b, 2026-10-08: reference rating 1,452.4. Published confidence bounds 1,444.9–1,459.8. 6,145 votes. Published gemma-4-31b configuration reference. Reasoning mode is not disclosed by the Arena dataset. Tested precision, runtime and sampling settings are not disclosed in this dataset. Selected Q8_0 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: 34.6 GB; Weights + KV cache + runtime overhead
Estimated generation: 18–27 tok/s
Estimated first token: ~4.1s

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 31B quantizations on NVIDIA L40S (48GB), estimated at 4K context. Weights and total memory are shown separately.
Quant	Weights	Total memory	Speed	Fit
Q2_K	11.8 GB	13.7 GB	43–65 tok/s	Runs great
Q3_K_M	14.7 GB	16.7 GB	36–54 tok/s	Runs great
Q4_K_M	18.3 GB	20.3 GB	30–45 tok/s	Runs great
Q5_K_M	21.7 GB	23.6 GB	26–39 tok/s	Runs great
Q6_K	25.2 GB	27.2 GB	23–34 tok/s	Runs great
Q8_0	32.6 GB	34.6 GB	18–27 tok/s	Runs great
FP16	61.4 GB	63.4 GB	— not resident	Won't fit

Open the full calculator →Everything a NVIDIA L40S (48GB) runs Gemma 4 31B on other machines What to buy to run it →

Ran Gemma 4 31B on a NVIDIA L40S (48GB)? Add your real numbers →