Dedicated GPU instances (Compute) are billed per-hour at fixed rates by instance type. Unlike Serverless, Compute instances run continuously regardless of utilization. Pricing is configured per instance type and visible in the dashboard.
Instance Type
Description
1xH100-SXM
Single H100 GPU, 16 vCPU, 200GB RAM, 1TB SSD
8xH100-SXM
Eight H100 GPUs, 128 vCPU, 1600GB RAM, 8TB SSD
Contact the fal team or check the dashboard for current per-hour rates.