Can L40S 48GB run gemma-3-27b-it?

NVIDIA · 48 GB GDDR6 · 864 GB/s bandwidth · Google · 27.4B · model ctx up to 128K

Yes — the L40S 48GB (48 GB GDDR6) can run gemma-3-27b-it. At Q4 with the default 8K context it needs ≈ 21.51 GB VRAM (weights 15.07 GB + KV cache 5.44 GB + 1 GB runtime overhead), which fits within 48 GB. Decode at Q4/8K: ≈ 20.2 tok/s (estimated, batch 1). FP16 needs ≈ 61.24 GB — does not fit on this card. Estimates come from memory-bandwidth math; rows tagged measured override estimates. Relative ranking is more reliable than absolute tok/s. Methodology.

Fit & speed by quant and context

Quant	Context	VRAM needed	Fits?	tok/s (decode)
Q4	4K	18.79 GB	Yes	20.2 est.
Q4	8K default	21.51 GB	Yes	20.2 est.
Q4	32K	37.83 GB	Yes	20.2 est.
FP16	4K	58.52 GB	No	10.7 est.
FP16	8K default	61.24 GB	No	9.8 est.
FP16	32K	77.56 GB	No	6.1 est.

Fits = weights + KV(ctx) + 1 GB overhead ≤ 95% of VRAM. Measured anchors are context-agnostic; the fit verdict is recomputed per context. A missing tok/s means the model is far beyond this card (offload-only territory).

VRAM breakdown at 8K context

Quant	Weights	KV cache	Overhead	Total needed	L40S 48GB VRAM
Q4	15.07 GB	5.44 GB	1 GB	21.51 GB	48 GB
FP16	54.8 GB	5.44 GB	1 GB	61.24 GB	48 GB

Other GPUs that run gemma-3-27b-it

• B200 192GB SXM — ≈187.2 tok/s Q4 est.
• MI300X 192GB — ≈124.0 tok/s Q4 est.
• H200 141GB SXM5 — ≈112.3 tok/s Q4 est.
• H100 80GB SXM5 — ≈78.4 tok/s Q4 est.
• Gaudi 3 128GB — ≈76.6 tok/s Q4 est.
• A100 80GB SXM4 — ≈47.7 tok/s Q4 est.
• RTX 5090 32GB — ≈41.9 tok/s Q4 est.
• RTX 4090 24GB — ≈23.6 tok/s Q4 est.

All GPUs for gemma-3-27b-it →

Other models for the L40S 48GB

• Qwen3-0.6B — ≈687.3 tok/s Q4 est.
• gemma-3-1b-it — ≈549.8 tok/s Q4 est.
• Llama-3.2-1B-Instruct — ≈458.2 tok/s Q4 est.
• Qwen3-1.7B — ≈274.9 tok/s Q4 est.
• Qwen3-30B-A3B — ≈183.3 tok/s Q4 est.
• Qwen3-Coder-30B-A3B-Instruct — ≈183.3 tok/s Q4 est.
• GLM-4.7-Flash — ≈183.3 tok/s Q4 est.
• Qwen3-30B-A3B-Instruct-2507 — ≈183.3 tok/s Q4 est.

All models on the L40S 48GB →