Can L40S 48GB run Qwen3-32B?

NVIDIA · 48 GB GDDR6 · 864 GB/s bandwidth · Alibaba · 32.8B · model ctx up to 128K

Yes — the L40S 48GB (48 GB GDDR6) can run Qwen3-32B. At Q4 with the default 8K context it needs ≈ 25.6 GB VRAM (weights 18.04 GB + KV cache 6.56 GB + 1 GB runtime overhead), which fits within 48 GB. Decode at Q4/8K: ≈ 16.8 tok/s (estimated, batch 1). FP16 needs ≈ 73.16 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	22.32 GB	Yes	16.8 est.
Q4	8K default	25.6 GB	Yes	16.8 est.
Q4	32K	45.28 GB	Yes	16.8 est.
FP16	4K	69.88 GB	No	6.2 est.
FP16	8K default	73.16 GB	No	5.7 est.
FP16	32K	92.84 GB	No	3.5 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	18.04 GB	6.56 GB	1 GB	25.6 GB	48 GB
FP16	65.6 GB	6.56 GB	1 GB	73.16 GB	48 GB

Other GPUs that run Qwen3-32B

• B200 192GB SXM — ≈155.2 tok/s Q4 est.
• MI300X 192GB — ≈102.8 tok/s Q4 est.
• H200 141GB SXM5 — ≈93.1 tok/s Q4 est.
• H100 80GB SXM5 — ≈65.0 tok/s Q4 est.
• Gaudi 3 128GB — ≈63.6 tok/s Q4 est.
• A100 80GB SXM4 — ≈39.6 tok/s Q4 est.
• RTX 5090 32GB — ≈34.8 tok/s Q4 est.
• RTX 6000 Ada 48GB — ≈18.6 tok/s Q4 est.

All GPUs for Qwen3-32B →

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 →