Pricing

Simple, usage-based pricing

Pick a plan, pay only for what you use. No upfront commitments.

For individual researchers and small projects.

Free to start

Pay only for compute

For teams and production AI workloads.

$99/mo

+ compute usage

For large teams with dedicated infrastructure.

Custom pricing

Volume discounts available

GPU pricing

All prices in USD per GPU per hour. Spot prices update every 5 minutes.

GPU	VRAM	TFLOPS	On-demand	Spot	Best for
H100 SXM5	80 GB HBM3	989	$3.29	$1.18	LLM Training, Large-scale AI
H100 PCIe	80 GB HBM2e	756	$2.49	$0.89	Inference, Fine-tuning
A100 80GB	80 GB HBM2e	312	$1.89	$0.68	Training, Inference, HPC
L40S	48 GB GDDR6	362	$1.29	$0.46	Rendering, Inference, Dev
A100 40GB	40 GB HBM2	312	$1.19	$0.43	Training, Research
RTX 4090	24 GB GDDR6X	165	$0.69	$0.25	Dev / Test, Small Models, Rendering