NVIDIA L40S — Local LLM Performance & Compatibility

Written by Jakub Rusinowski · Last updated July 21, 2026

The mainstream rack-server inference card of the Ada generation: 48 GB GDDR6, passive cooling (needs server chassis airflow), widely available in 1–8× configurations from Dell, HPE, Lenovo, and Supermicro. A common building block for on-prem AI pilots because it slots into standard 2U servers without exotic power or cooling.

Technical Specifications

VRAM48 GB
Memory Bandwidth864 GB/s
TDP350 W
ArchitectureAda Lovelace AD102
Release Year2023
MSRP at Launch$7,499
Inference Speed (Llama 3.1 8B Q4_K_M)77–147 tok/s (estimated)
Inference Speed (Llama 3.3 70B Q4_K_M)12–23 tok/s (estimated)
Buy This HardwareNVIDIA L40S 48GB — 48 GB VRAM · 350 W board powerDeploy in the Cloud NowNVIDIA L40S on RunPod — from $0.86/hr · rate checked 2026-08

or compare on Vast.ai

As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.

Affiliate disclosure: Some links on this page are affiliate links — if you buy through them, LLM Configurator may earn a commission at no extra cost to you. As an Amazon Associate, LLM Configurator earns from qualifying purchases.
NVIDIA L40S 48GB
48 GB VRAM · 350 W board power
2026 prices are volatile — check the current listing.
Check price on Amazon

LLMs Compatible with 48 GB VRAM

All models below run comfortably in 48 GB VRAM with Q4_K_M quantization.

Llama 3.3Llama 3.3 70B Instruct · 43 GB VRAM · Q4_K_M · ollama run llama3.3
Llama 3.1 FamilyLlama 3.1 8B Instruct · 6 GB VRAM · Q4_K_M · ollama run llama3.1
DeepSeek R1DeepSeek R1 Distill Qwen 32B · 20 GB VRAM · Q4_K_M · ollama run deepseek-r1:32b
Qwen 3Qwen 3 32B · 21 GB VRAM · Q4_K_M · ollama run qwen3:32b
Qwen 3.5Qwen 3.5 35B-A3B · 22 GB VRAM · Q4_K_M · ollama run qwen3.5:35b-a3b
Qwen 3.6Qwen 3.6 35B-A3B · 22 GB VRAM · Q4_K_M · ollama run qwen3.6:35b-a3b
Gemma 3Gemma 3 27B Instruct · 17 GB VRAM · Q4_K_M · ollama run gemma3:27b
Gemma 4Gemma 4 31B · 20 GB VRAM · Q4_K_M · ollama run gemma4:31b

Best Use Cases

Quick Start with Ollama

Install Ollama then run the recommended model for this GPU:

ollama run llama3.3:70b

FAQ

Can the NVIDIA L40S run local LLMs?

Yes — the NVIDIA L40S has 48 GB VRAM and runs The mainstream rack-server inference card of the Ada generation: 48 GB GDDR6, passive cooling (needs server chassis airf

How fast is the NVIDIA L40S for AI inference?

The NVIDIA L40S is estimated to run Llama 3.1 8B at 77–147 tok/s with Q4_K_M quantization. For Llama 3.3 70B the estimate is 12–23 tok/s. These are modelled estimates, not measurements — see /en/methodology.

What LLMs can I run on 48 GB VRAM?

With 48 GB you can run: Llama 3.3, Llama 3.1 Family, DeepSeek R1, Qwen 3, Qwen 3.5. Use Ollama for the easiest setup: ollama run llama3.3:70b.

Can I Run It? — NVIDIA L40S

Compare Similar GPUs

VRAM Tier

Buying Guide

← All GPU Reviews | Check Your Hardware | Full Benchmarks | Can I Run It?