Is fal.ai free? Credits, keys and what you pay for

Not as a standing free tier. fal's pricing docs describe a prepaid credit model and don't list a free tier: you buy credits in advance, and each model call is billed by its output, such as per image, per megapixel, per second of video or per video. You pay only for successful outputs, and fal says you are never charged for server errors or time spent waiting in the queue. Creating an API key costs nothing.

These facts are quoted from fal's pricing docs, FAQ, concurrency limits and key-based authentication pages, read on 2026-09-29. The FAQ does mention free credits and coupons (below). Per-model rates change, so read each model's page for its price.

How does fal bill model calls?

"fal uses a prepaid credit model. You purchase credits in advance and they are drawn down as you use the platform." Credits fund both Model API usage and your concurrency limit. The billing unit depends on the model:

• Models without a fixed per-output price fall back to per-second billing by GPU machine type.
• A request that fails with HTTP 500 or higher costs nothing.
• Queue time before a runner starts is free; "Only the actual inference work counts toward your bill."
• You can read a model's unit price through fal's Platform APIs before you run it.

Does fal give free credits?

Sometimes, but the docs don't say who gets them. The FAQ, answering "Do my credits expire?", says purchased credits expire 365 days from purchase, and "Free credits and coupons have variable expiration depending on the specific grant, ranging from 1 week to 1 year." It doesn't describe a sign-up grant, so don't plan on one.

When the balance drops below your account's lock threshold, the account is locked and API requests are rejected until you add credits. Enterprise customers on invoice billing are not locked automatically.

Credits also set how many requests run at once: fal says every new account starts at 2 concurrent requests, rising as you buy credits. fal.ai API rate limit and concurrency covers the limits.

Is a fal API key free?

Yes. You generate keys from the fal dashboard at fal.ai/dashboard/keys, and the key page lists no fee for one. Keys have two scopes, ADMIN and API; fal recommends API if you only call ready-made models. The calls a key makes draw on your credits. For wiring a key into an agent, see fal MCP server setup and billing.

What should I check when comparing fal with other hosts?

Check what each host charges for a failed job and what raises its concurrency. On Sume, for example, the generation admission docs say failed jobs release or refund their reserved cost, and concurrency is "plan-only": prepaid top-ups do not raise it. For how fal and Replicate bill, see fal vs Replicate.