Can L40S 48GB run gpt-oss-120b?

NVIDIA · 48 GB GDDR6 · 864 GB/s bandwidth · OpenAI · 120.4B (active 5.1B) · MoE · model ctx up to 128K

No — gpt-oss-120b at Q4 needs ≈ 67.73 GB even at 4K context, beyond the L40S 48GB's 48 GB GDDR6. It would only run with heavy CPU/disk offload. Decode at Q4/8K: ≈ 152.4 tok/s (estimated, batch 1). FP16 needs ≈ 242.82 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	67.73 GB	No	154.7 est.
Q4	8K default	68.24 GB	No	152.4 est.
Q4	32K	71.3 GB	No	139.6 est.
FP16	4K	242.31 GB	No	—
FP16	8K default	242.82 GB	No	—
FP16	32K	245.88 GB	No	—

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	66.22 GB	1.02 GB	1 GB	68.24 GB	48 GB
FP16	240.8 GB	1.02 GB	1 GB	242.82 GB	48 GB

Other GPUs that run gpt-oss-120b

• B200 192GB SXM — ≈998.2 tok/s Q4 est.
• MI300X 192GB — ≈661.3 tok/s Q4 est.
• H200 141GB SXM5 — ≈598.9 tok/s Q4 est.
• H100 80GB SXM5 — ≈418.0 tok/s Q4 est.
• Gaudi 3 128GB — ≈408.8 tok/s Q4 est.
• A100 80GB SXM4 — ≈254.4 tok/s Q4 est.
• Mac Studio M3 Ultra 96GB — ≈102.2 tok/s Q4 est.
• Mac Studio M2 Ultra 192GB — ≈79.7 tok/s Q4 measured

All GPUs for gpt-oss-120b →

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 →