fal.ai

What's good about fal.ai

• Strong catalog of generative media models behind a single API
• Per-second billing for serverless GPU deployments
• Specialized inference optimizations for diffusion and audio

fal.ai GPU prices

6 GPU models across 6 configurations. Rank compares fal.ai's cheapest on-demand listing for each GPU with every provider's cheapest.

What a run costs on fal.ai

24 hours at today's cheapest listing for each GPU, at the size it is listed.

• 1× H100 SXM for 24 hours

  At $4.50/GPU-hr

  $108.00

  $31.20 over the market median rate

• 1× H200 for 24 hours

  At $6.00/GPU-hr

  $144.00

  $35.05 over the market median rate

• 1× RTX PRO 6000 for 24 hours

  At $4.00/GPU-hr

  $96.00

  $44.52 over the market median rate

The market median is every provider's cheapest on-demand price for that GPU.

Billing & terms

Output-Based Model API Pricing

Hosted model endpoints billed per generated output unit — per image, per megapixel, per second of video or per video

Per-Second GPU Pricing

Custom deployments billed per second of GPU runtime, with scale-to-zero

Enterprise Contracts

Volume commitments and dedicated capacity for high-throughput customers, with discounted GPU rates below the published list price

Regions

Multi-region serverless infrastructure

Support

Documentation, community channels and enterprise support for paid customers

Watch out for

• Less suited to long-running fixed-instance training
• Custom GPU deployments require contacting support to get started
• Lower-tier consumer GPUs are not part of the catalog

Features

Hosted Model Catalog

Production endpoints for image, video and audio models billed by output unit (per image, per megapixel, per second or per video)

Custom GPU Deployments

Run private models on dedicated NVIDIA GPUs with autoscaling and scale-to-zero

Optimized Runtimes

Inference engines tuned for diffusion and audio workloads

Getting started

1. Step 1

   Create an account

   Sign up and generate an API key

2. Step 2

   Pick a hosted model or upload your own

   Choose from the catalog or define a custom GPU-backed deployment

3. Step 3

   Call the API

   Invoke endpoints from any language using the REST or SDK clients