Find the cheapest cloud GPU.
65 providers, 67 GPUs, 1,841 published prices. Type a card, get the cheapest — here or from your coding agent.
Updated
Popular models
RTX A4000 16GB$0.09
Vast.aiRTX 3080 10GB$0.11 Vast.aiRTX 3080 Ti 12GB$0.13 Vast.aiRTX 3090 24GB$0.15 Vast.aiRTX 5060 Ti 16GB$0.15O Omega GradientRTX 4080 16GB$0.15A Akash NetworkTesla V100 32GB$0.16
GeoddRTX A5000 24GB$0.16
RunPod Community CloudRTX 5070 Ti 16GB$0.16 Vast.aiRTX 4060 Ti 16GB$0.17O Omega GradientRTX 4070 12GB$0.18 Vast.aiRTX 4070 Super 12GB$0.18 Vast.aiRTX 5070 12GB$0.19 Vast.aiRTX 3090 Ti 24GB$0.22 Vast.aiTesla V100 16GB$0.23
VerdaRTX PRO 4000 Blackwell 24GB$0.24 Vast.aiRTX 4080 Super 32GB$0.25
GPUhubRTX 5080 16GB$0.26 Vast.aiRTX 4000 Ada 20GB$0.29A Akash NetworkA30 24GB≈$0.33S SeewebRTX A6000 48GB$0.33 RunPod Community CloudRTX 4090 24GB$0.34A Akash NetworkRTX 5090 32GB$0.35H HyperAIA40 48GB$0.35 RunPod Community CloudRTX PRO 4500 Blackwell 32GB$0.35 Vast.aiA100 SXM4 40GB$0.38 Vast.aiRTX PRO 4000 Blackwell SFF 24GB≈$0.39
HetznerTesla T4 16GB$0.42
Amazon Web ServicesL4 24GB≈$0.43S SeewebRTX 4090 48GB$0.44 GPUhubL40 48GB$0.49 Vast.aiRTX 5000 Ada 32GB$0.50O OpenRelayA100 PCIe 40GB$0.54 Vast.aiRTX 6000 Ada 48GB$0.54 Vast.aiRTX PRO 5000 Blackwell 48GB$0.59 Vast.aiL40S 48GB$0.60 Vast.aiQuadro RTX 6000 24GB$0.69
LambdaGB10 DGX Spark 128GB$0.75
EnvergeRTX PRO 6000 Blackwell 96GB$0.80H HyperAIA100 PCIe 80GB$0.80 Vast.aiA100 SXM4 80GB$0.80 Vast.aiRTX PRO 6000D 84GB$0.97 GPUhubA10G 24GB$1.01 Amazon Web ServicesA10 24GB≈$1.01
OVHcloudTesla V100S 32GB≈$1.01 OVHcloudGaudi 2 96GB$1.03
Cyfuture AIA100 80GB$1.07
GPU.aiA800 80GB$1.18 GPUhubH100 SXM5 80GB$1.74 Vast.aiH100 PCIe 80GB$1.80H HyperAIInstinct MI300X 192GB$1.99O OpenRelayH200 SXM5 141GB$2.25
SesterceGH200 96GB$2.29 LambdaH100 NVL 94GB$2.59 RunPod Community CloudInstinct MI325X 256GB$3.04 Cyfuture AIH200 NVL 141GB$3.39 GeoddB200 SXM6 180GB$4.50 SesterceB300 SXM6 288GB$6.60
Massed ComputeInstinct MI355X 288GB$8.60O Oracle Cloud InfrastructureGB200 NVL72 192GB$9.99
FalGB300 NVL72 288GB$11.82 Verda
Vast.aiRTX 3080 10GB$0.11 Vast.aiRTX 3080 Ti 12GB$0.13 Vast.aiRTX 3090 24GB$0.15 Vast.aiRTX 5060 Ti 16GB$0.15O Omega GradientRTX 4080 16GB$0.15A Akash NetworkTesla V100 32GB$0.16 GeoddRTX A5000 24GB$0.16 RunPod Community CloudRTX 5070 Ti 16GB$0.16 Vast.aiRTX 4060 Ti 16GB$0.17O Omega GradientRTX 4070 12GB$0.18 Vast.aiRTX 4070 Super 12GB$0.18 Vast.aiRTX 5070 12GB$0.19 Vast.aiRTX 3090 Ti 24GB$0.22 Vast.aiTesla V100 16GB$0.23 VerdaRTX PRO 4000 Blackwell 24GB$0.24 Vast.aiRTX 4080 Super 32GB$0.25 GPUhubRTX 5080 16GB$0.26 Vast.aiRTX 4000 Ada 20GB$0.29A Akash NetworkA30 24GB≈$0.33S SeewebRTX A6000 48GB$0.33 RunPod Community CloudRTX 4090 24GB$0.34A Akash NetworkRTX 5090 32GB$0.35H HyperAIA40 48GB$0.35 RunPod Community CloudRTX PRO 4500 Blackwell 32GB$0.35 Vast.aiA100 SXM4 40GB$0.38 Vast.aiRTX PRO 4000 Blackwell SFF 24GB≈$0.39 HetznerTesla T4 16GB$0.42 Amazon Web ServicesL4 24GB≈$0.43S SeewebRTX 4090 48GB$0.44 GPUhubL40 48GB$0.49 Vast.aiRTX 5000 Ada 32GB$0.50O OpenRelayA100 PCIe 40GB$0.54 Vast.aiRTX 6000 Ada 48GB$0.54 Vast.aiRTX PRO 5000 Blackwell 48GB$0.59 Vast.aiL40S 48GB$0.60 Vast.aiQuadro RTX 6000 24GB$0.69 LambdaGB10 DGX Spark 128GB$0.75 EnvergeRTX PRO 6000 Blackwell 96GB$0.80H HyperAIA100 PCIe 80GB$0.80 Vast.aiA100 SXM4 80GB$0.80 Vast.aiRTX PRO 6000D 84GB$0.97 GPUhubA10G 24GB$1.01 Amazon Web ServicesA10 24GB≈$1.01 OVHcloudTesla V100S 32GB≈$1.01 OVHcloudGaudi 2 96GB$1.03 Cyfuture AIA100 80GB$1.07 GPU.aiA800 80GB$1.18 GPUhubH100 SXM5 80GB$1.74 Vast.aiH100 PCIe 80GB$1.80H HyperAIInstinct MI300X 192GB$1.99O OpenRelayH200 SXM5 141GB$2.25 SesterceGH200 96GB$2.29 LambdaH100 NVL 94GB$2.59 RunPod Community CloudInstinct MI325X 256GB$3.04 Cyfuture AIH200 NVL 141GB$3.39 GeoddB200 SXM6 180GB$4.50 SesterceB300 SXM6 288GB$6.60 Massed ComputeInstinct MI355X 288GB$8.60O Oracle Cloud InfrastructureGB200 NVL72 192GB$9.99 FalGB300 NVL72 288GB$11.82 Verdacheapest on-demand rate per GPU-hour, 61 cards, right now
Head to head
Two cards, every published spec and every provider's price for both.
H100 vs A100Same 80 GB of memory, two generations apart. What separates them is bandwidth, FP8, and the rate each provider is charging today.H200 vs H100Same Hopper silicon, 141 GB of memory instead of 80. The published compute figures are identical; what you are paying for is capacity, bandwidth, and fewer cards per model.B200 vs H100Blackwell against Hopper: more than twice the memory and bandwidth, and an FP4 path the H100 does not have. What the published figures say, and what providers are charging for each today.B200 vs H200The two largest-memory NVIDIA cards a provider will rent you. Closer on capacity than the names suggest; the gap is bandwidth, compute, FP4 — and price.L40S vs A100Ada workstation silicon against Ampere datacenter silicon. The L40S has FP8 and the newer architecture; the A100 has 80 GB of HBM and the bandwidth to feed it. Which one a job needs, and what each costs to rent today.
Provider reviews
How each one bills, what it rents and what it guarantees — from its own site.
ToolFind the cheapest GPU for the job.Paste a Hugging Face model. See which cards it fits on at each precision, and what the cheapest of them costs.Open the recommender →ToolWhat will a whole run cost?Hours, cards, provider — the total for the job rather than the hourly rate, from the same published prices.Open the calculator →
65 providers, read every six hours
Published self-serve list prices only. Anyone behind a login or a sales form is excluded, with the reason stated.
