NVIDIA L40S vs NVIDIA RTX A6000
What each costs to rent by the hour across the providers we track — NVIDIA L40S on 44 providers, NVIDIA RTX A6000 on 23.
As of , the NVIDIA L40S has a median rate card of $1.42 per GPU-hour versus $0.59 for the NVIDIA RTX A6000 — $0.83 (58%) more.
Side by side
| NVIDIA L40S | NVIDIA RTX A6000 | |
|---|---|---|
| Architecture | Ada | Ampere |
| VRAM | 48 GB | 48 GB |
| Form factor | PCIe | PCIe |
| Median rate card, per GPU-hour | $1.42 | $0.59 |
| Cheapest listed rate | $0.38 Lium, community, floor captured | $0.27 Yotta Labs, spot captured |
| Median per GB of VRAM, per hour | $0.030 | $0.012 |
| Providers listing it | 44 | 23 |
Providers that list both
| Provider | NVIDIA L40S | NVIDIA RTX A6000 | RTX A6000 vs L40S |
|---|---|---|---|
 GPU.ai | $0.80 from captured | $0.40 from captured | −$0.40 (−50%) |
 Airon | $0.90 from captured | $0.60 from captured | −$0.30 (−33%) |
 Seeweb | $0.96 captured | $0.83 captured | −$0.12 (−13%) |
 Massed Compute | $0.97 captured | $0.55 captured | −$0.42 (−43%) |
 Sesterce | $0.97 from captured | $0.55 from captured | −$0.42 (−43%) |
 HexGrid | $1.00 Secure Cloud captured | $0.55 Secure Cloud captured | −$0.45 (−45%) |
 RunPod | $1.09 Secure Cloud captured | $0.59 Secure Cloud captured | −$0.50 (−46%) |
 QuantaCloud | $1.09 Midwest - us-midwest-2 captured | $0.48 Midwest - us-midwest-2 captured | −$0.61 (−56%) |
 Koyeb | $1.20 captured | $0.75 captured | −$0.45 (−37%) |
 EmpirioLabs | $1.80 captured | $0.90 captured | −$0.90 (−50%) |
 Verda | $1.94 captured | $0.84 captured | −$1.09 (−56%) |
Medians are median rate cards — each provider once, at its cheapest on-demand rate, community capacity and marketplace floors out — the same figure as on each card's page; the cheapest listed rate may be spot or a community tier and is labelled as such. Compare providers directly on the overview, or see each GPU's full table: NVIDIA L40S, NVIDIA RTX A6000. On your own page: embed this comparison (/embed/gpus/l40s-vs-rtx-a6000, an iframe, no script).
Data updated: