interactive series
Subscriptions
Start with 200 trial credits for three days, then continue on the Basic monthly subscription. Paid plans can still be selected monthly or annually.
3-day trial
Try the API with 200 credits for three days, then continue on Basic monthly.
200 credits for three days, then Basic monthly
Starter API work
For prototypes, product tests, and short creator workflows.
$19.20/month billed monthly
Most teams
Most usefulFor launch campaigns, batch testing, and repeatable API usage.
$64/month billed monthly
High-volume studio
For agencies, studios, and scheduled video operations.
$192/month billed monthly
Each paid plan is a recurring subscription. Credits reset at the start of each billing period. Standard 768P reserves 10 credits per second; Turbo 768P reserves 5 credits per second.
Source-backed guide
Four API plans with monthly and annual subscriptions, credits, and source-backed model-rate context.
This guide covers video pricing.
Read first
For video pricing, the safest public number is the dated fal model-page snapshot checked on September 4, 2026. fal lists text-to-video and image-to-video launch rates at $0.0125/sec for 480P and $0.02/sec for 768P. The same model pages state that the listed launch discount ends on September 7, 2026, after which the visible list rates are $0.05/sec and $0.08/sec. Source: Official fal model source.
That means a quick five-second 768P video check is priced very differently during the launch window than it is under the post-discount list rate. The page should never treat the promotion as permanent. For a paid workflow, check the live fal model page and the fal pricing documentation again before creating a customer-facing quote or internal budget. Source: Official fal billing source.
Source table
All rows are copied from fal model or billing pages and are labeled as a snapshot, not a contract.
| Route family | Resolution | Visible rate | Budget note | Source |
|---|---|---|---|---|
| text or image | 480P | $0.0125/sec launch rate | $0.05/sec after the listed launch discount | Official fal model source |
| text or image | 768P | $0.02/sec launch rate | $0.08/sec after the listed launch discount | Official fal model source |
| Turbo text or image | 480P | $0.00625/sec launch rate | $0.025/sec after the listed launch discount | Official fal model source |
| Turbo text or image | 768P | $0.01/sec launch rate | $0.04/sec after the listed launch discount | Official fal model source |
| reference video output | 480P | $0.05/sec output video | Reference tokens may add cost after the first 4096 tokens | Official fal model source |
| reference video output | 768P | $0.08/sec output video | Reference tokens may add cost after the first 4096 tokens | Official fal model source |
Source table
These examples multiply seconds by the fal-visible per-second rates. They do not include reference-token charges.
| Scenario | Launch-rate estimate | Post-discount list estimate | Reason to run it | Source |
|---|---|---|---|---|
| 5 sec video at 480P | $0.0625 | $0.25 | Fast prompt screening before spending on detail. | Official fal model source |
| 5 sec video at 768P | $0.10 | $0.40 | Short quality check at the higher video resolution. | Official fal model source |
| 15 sec video at 768P | $0.30 | $1.20 | Maximum-length draft in fal comparison framing. | Official fal comparison source |
| 10 sec Turbo at 768P | $0.10 | $0.40 | Throughput-oriented variant when speed and lower cost matter. | Official fal model source |
Practical reading
The largest pricing mistakes usually come from mixing route families or forgetting that reference mode has more than output-video seconds.
For text-to-video and image-to-video, start with seconds multiplied by the visible rate for 480P or 768P. Keep the date beside the number so nobody carries a launch-rate assumption into a later budget.
fal reference-to-video pricing says the first 4096 reference tokens are free, then additional reference tokens are billed. Image, video, and audio references have different token behavior, so reference-mode budgets need a separate line item.
fal pricing documentation says billing is based on generated output and that queue waiting is not charged. That helps separate infrastructure latency from generated seconds when reading logs.
Official media
This fal-hosted sample is included so the page has a concrete public video reference beside the pricing or route facts.
Official fal sample output from the Text to Video Max model page.
Source: fal Text to Video Max model page
FAQ
fal lists the Standard model at $0.02/sec for 768P during the launch-rate window checked on September 4, 2026.
Use the visible post-discount list rate on fal: $0.08/sec for 768P and $0.05/sec for 480P, then re-check fal before paid use.
No. It includes output video seconds and can add reference-token charges after the free token allowance.
Source map
Every hard claim on this page is tied to official fal material or an independent benchmark source. Source basis checked September 4, 2026.
Used for public API tier names, subscription periods, prices, credit allowances, and checkout routing.
Used for the text-to-video route, launch-rate pricing, sample output video, and public model-page positioning.
Used for image-to-video positioning, first-frame and end-frame examples, current rates, and sample output media.
Used for reference-to-video pricing, token rules, and sample output media.
Used for the rule that model billing depends on generated output, that failed server jobs and queue waiting are not charged, and that live model pages should be checked before paid use.