🚀 One API for every frontier model on FastInfra

Browse models

Transparent per-token pricing → compare providers on FastInfra

View pricing

OpenAI-compatible chat completions → start in minutes

Read the docs

⚡ Server Auction → Enterprise bare metal from $78.15/mo (173 in stock)

Browse deals

GPU Cloud · Billed per minute

GPU cloud pods billed per minute: price yours before you buy

FastInfra GPU Cloud rents dedicated GPU pods from RTX 4090 to H200, billed per minute for exactly as long as the pod exists. The public configurator shows the live hourly and monthly price for the GPU type, count, region, container template and attached storage before anything is provisioned, so you can size a training run or an inference deployment and know its cost up front. Pods offer SSH and JupyterLab at deploy time plus an optional persistent volume that survives stops.

Pick a GPU, size the disks, and see exactly what it costs per minute, per day, and per month. Nothing is provisioned and nothing is charged until you deploy — configure freely.

Any region — we pick the best available datacenter when you deploy.

1 · Choose a GPU

2 · Size your pod

Nothing is reserved while you configure
1

20 GB

Holds your image, CUDA cache, and scratch files. Ephemeral — wiped when the pod stops.

None

Mounted at /workspace and survives stops. Datasets and checkpoints belong here.

$
lasts —

How long this configuration can run on that much credit, assuming it stays running the whole time.

What fits in — of VRAM

Rule-of-thumb weight sizing
FP16 / BF16
—
2 bytes per parameter
8-bit
—
1 byte per parameter
4-bit
—
0.5 bytes per parameter

Typically used for: —

Estimates reserve 20% of VRAM for the KV cache and activations. Real headroom depends on context length, batch size, and your serving stack — treat these as a starting point, not a guarantee.

—

—

GPU compute —
Storage while running —
Total while running —
While stopped (volume only) —
1 hour
—
1 day
—
1 week
—
1 month
—
Create account to deploy →

Minimum credit to start this pod: — (1h of runtime). Compute billing stops the moment you stop the pod.

Billed by the minute

Charges are drawn from your prepaid credits every minute the pod is provisioning or running. Stop it and compute charges end instantly — no minimum term, no reservation fees.

Secure vs value tier

Secure runs in vetted datacenters with the highest reliability and availability. Value uses lower-cost community hosts — same GPU model, cheaper rate, more variable capacity.

Storage that persists

A persistent volume survives stops and restarts. A stopped pod only pays the volume storage rate ($0.270/GB/month) — never for the GPU.

Transparent pricing

Every rate shown includes our 35% platform markup on underlying compute and storage — there is nothing added at checkout. Up to 2 active pods per account by default.

What you get on every pod

Your container, your rules

Any public Docker image — PyTorch, TensorFlow, vLLM, ComfyUI, Ollama, or your own build. Set environment variables and expose up to 10 ports.

Direct access

Enable SSH and JupyterLab at deploy time. TCP ports get a public IP and port you connect to directly; HTTP ports are proxied for you.

One wallet, one dashboard

GPU pods draw from the same FastInfra credits as API inference. Live usage, per-minute charges, and runtime totals appear on each pod's detail page.

Cloud GPU prices per hour

Every GPU you can deploy today, with its VRAM and lowest on-demand price per GPU-hour. Open a GPU for specs, multi-GPU pricing and which models fit.

GPUVRAMArchitectureFrom / hrSpot / hr
MI350X 294 GB — $7.41 —
B300 288 GB Blackwell $9.37 —
MI300X 192 GB CDNA 3 $3.23 —
B200 180 GB Blackwell $8.07 —
H200 NVL 143 GB Hopper $5.12 —
H200 SXM 141 GB Hopper $4.85 —
RTX PRO 6000 MaxQ 96 GB Blackwell $2.21 —
RTX PRO 6000 96 GB Blackwell $2.28 —
RTX PRO 6000 WK 96 GB Blackwell $2.28 —
H100 NVL 94 GB Hopper $3.50 —
A100 PCIe 80 GB Ampere $1.61 —
A100 SXM 80 GB Ampere $1.88 —
H100 PCIe 80 GB Hopper $2.69 —
H100 SXM 80 GB Hopper $3.63 —
RTX A6000 48 GB Ampere $0.446 —
A40 48 GB Ampere $0.797 —
L40 48 GB Ada Lovelace $0.932 —
RTX 6000 Ada 48 GB Ada Lovelace $0.999 —
L40S 48 GB Ada Lovelace $1.07 —
RTX PRO 5000 48 GB Blackwell $1.11 —
PRO 6000 MIG 48GB 48 GB Blackwell $1.88 —
A100 SXM 40GB 40 GB Ampere $1.35 —
RTX 5000 Ada 32 GB Ada Lovelace $0.662 —
RTX 5090 32 GB Blackwell $0.932 —
RTX PRO 4500 32 GB Blackwell $0.972 —
RTX PRO 4500 SE 32 GB Blackwell $0.972 —
RTX A5000 24 GB Ampere $0.216 —
RTX 3090 24 GB Ampere $0.297 —
RTX 3090 Ti 24 GB Ampere $0.365 —
RTX 4090 24 GB Ada Lovelace $0.459 —
RTX PRO 4000 24 GB Blackwell $0.77 —
L4 24 GB Ada Lovelace $0.932 —
PRO 6000 MIG 24GB 24 GB Blackwell $0.932 —
RTX A4500 20 GB Ampere $0.257 —
RTX 4000 Ada 20 GB Ada Lovelace $0.27 —
RTX A4000 16 GB Ampere $0.23 —
V100 SXM2 16 GB Volta $0.311 —
RTX 2000 Ada 16 GB Ada Lovelace $0.324 —
RTX 4080 SUPER 16 GB Ada Lovelace $0.378 —
RTX 5080 16 GB Blackwell $0.527 —
RTX 3080 Ti 12 GB Ampere $0.243 —
RTX 4070 Ti 12 GB Ada Lovelace $0.257 —
RTX 3080 10 GB Ampere $0.23 —
RTX 3070 8 GB Ampere $0.176 —
RTX A2000 6 GB Ampere $0.162 —

Frequently asked questions

Which GPUs can I rent on FastInfra GPU Cloud?

RTX 4090, A100, H100 and H200 pods, plus other types as capacity appears. The configurator lists every available GPU with its VRAM and live hourly price.

How is a GPU pod billed?

Per minute while the pod exists, at the hourly rate shown in the configurator divided by 60. Attached volumes bill per GB-month and continue while a pod is stopped; delete the pod to stop all charges.

Do I need to deploy to see GPU prices?

No. The public configurator shows live prices for every GPU, count and region without an account. Deploying requires a signed-in account with enough prepaid balance for the configured minimum hours.

Can I run multiple GPUs in one pod?

Yes. Choose one to eight GPUs of the same type per pod, subject to the maximum shown for that GPU and cloud type; the hourly price scales with the GPU count.

What is the difference between GPU Cloud and the inference API?

GPU Cloud gives you a whole machine you control and pay for by the minute. The inference API is pay-per-token access to hosted models with no servers to manage. Many teams use the API for production traffic and GPU pods for fine-tuning or custom models.

Reviewed by Ashish Kumar, Head of Product · Updated October 15, 2026. Prices and counts on this page come from the live catalog.