fal.ai pricing
by falCoding tools
Custom / monthfal.ai Model APIs · billed monthly
fal charges per output with no subscription: for example FLUX.1 [schnell] is $0.003 a megapixel, Nano Banana is $0.0398 an image and Kling Video v3 Pro is $0.14 a second. Serverless GPUs for your own apps are listed from $4.50 an hour for an H100, with lower rates shown as low as $2.49 an hour.
Lowest price$4.50per hour · Serverless H100
Every plan, side by side
2 plans · prices in USD, before tax| Plan | Price | Usage | Seats | Best for |
|---|---|---|---|---|
| Model APIsRates differ by model | Custom | Per image, second or 1,000 characters | — | Generative media apps |
| Serverless H100List price; lower rates shown as low as $2.49 an hour | $4.50 per hour | — | — | Custom deployments |
Usage limitsOfficial
How much you actually get
Model APIsBilled by output unit per model: images, seconds of video or audio, 1,000 characters or 3D units
Serverless H100GPU list prices per hour: H100 $4.50, H200 $6.00, B200 $7.99, B300 $12.99
Official limits from the vendor's own pages.
Cost calculatorPick a plan to see its price and what it includes
What will you pay, and what do you get?
Plan
Custom
Priced by sales · contact fal.ai
What you get on fal.ai Model APIs
Generative media apps
Rates differ by model
- Pay per output with no monthly fee
- Images from $0.003 per megapixel (FLUX.1 schnell)
- Video from $0.015 to $0.14 per second, depending on model
- Audio from $0.0002 per second or $0.015 to $0.10 per 1,000 characters
- 3D models from $0.01 per credit or $0.02 per generation
UsagePer image, second or 1,000 characters
Official limitBilled by output unit per model: images, seconds of video or audio, 1,000 characters or 3D units
Similar tools, cheapest paid planCompare fal.ai with any tool →
fal.ai against the closest alternatives
fal.ai $4.50 per hour · Billed by output unit per model: images, seconds of video or audio, 1,000 characters or 3D unitsAssemblyAI $0.21 per hour · Streaming is billed on connection time, including idle time; multichannel audio is billed per channel; add-ons stack on the base rateDeepgram $0.0042 per minute · 150 concurrent streaming speech-to-text connections; Voice Agent and other limits apply per project and differ by plan and regionGitLab Duo $1 per credit · Code Suggestions: 50 per credit; Code Review Flow: 4 per credit; agentic chat and flows are billed per model call, e.g. 8 calls per credit on basic models, 2 on claude-sonnet-4.5 and 1.1 on claude-opus-4.6
Price changes · fal.aiNewest first
Every price and limit change
- No changes recorded yet.
Prices from fal.ai/pricing How we collect prices Some links are affiliate links; they never change the prices or the order shown. Disclosure