Nosana

An open-source GPU cloud built for AI and high-performance workloads

πŸ‡³πŸ‡±Haarlem, NLFounded2022Marketplace Β· Developer-firstUpdated
Cheapest / hr
$0.06
RTX 4060
Most expensive
$2.24
RTX PRO 6000
Average / GPU
$0.65
across 21 configs
GPU types
19
21 configurations
VRAM range
8–96GB
per single GPU
Regions
β€”
β€”

GPU pricing by type

Median price per GPU/hour across all configurations offered by Nosana.

Cheapest median$0.06
Most expensive$2.24
Configurations21
VRAM range8–96 GB

GPU configurations

21 offerings across 19 GPU types. Prices update every 4 hours.

Visit Nosana
21of 21
RegionTotal VRAMStorageStatus
8 GB VRAM per GPU
--Γ—18 GBβ€”β€”β€”$0.06$0.06
8 GB VRAM per GPU
--Γ—18 GBβ€”β€”β€”$0.08$0.08
10 GB VRAM per GPU
--Γ—110 GBβ€”β€”β€”$0.10$0.10
12 GB VRAM per GPU
--Γ—112 GBβ€”β€”β€”$0.10$0.10
16 GB VRAM per GPU
--Γ—116 GBβ€”β€”β€”$0.13$0.13
12 GB VRAM per GPU
--Γ—112 GBβ€”β€”β€”$0.15$0.15
16 GB VRAM per GPU
--Γ—116 GBβ€”β€”β€”$0.16$0.16
24 GB VRAM per GPU
--Γ—124 GBβ€”β€”β€”$0.19$0.19
16 GB VRAM per GPU
--Γ—116 GBβ€”β€”β€”$0.20$0.20
24 GB VRAM per GPU
--Γ—124 GBβ€”β€”β€”$0.20$0.20
48 GB VRAM per GPU
--Γ—148 GBβ€”β€”β€”$0.40$0.40
24 GB VRAM per GPU
--Γ—124 GBβ€”β€”β€”$0.40$0.40
48 GB VRAM per GPU
--Γ—148 GBβ€”β€”β€”$0.40$0.40
32 GB VRAM per GPU
--Γ—132 GBβ€”β€”β€”$0.50$0.50
40 GB VRAM per GPU
--Γ—140 GBβ€”β€”β€”$0.61$0.61
--Γ—148 GBβ€”β€”β€”$0.70$0.70
80 GB VRAM per GPU
--Γ—180 GBβ€”β€”β€”$0.90$0.90
NVIDIA H100Trending
80 GB VRAM per GPU
--Γ—180 GBβ€”β€”β€”$1.50$1.50
96 GB VRAM per GPUΒ· 3 configs
--Γ—6576 GBβ€”β€”β€”$2.20–$2.34$13.20

See incorrect data?

Frequently asked questions

What GPUs does Nosana offer?
Nosana offers 19 GPU types across 21 configurations, including RTX 3090, RTX 5070, RTX 4060, RTX 5090, RTX 5080, RTX 3070 and 13 more. The full list with live hourly pricing is shown above.
How much does Nosana cost?
Nosana GPU pricing starts at $0.06 per GPU-hour, with an average of $0.65 across its on-demand configurations. Spot and reserved options can be cheaper for interruptible or committed workloads.