# fal API pricing per model: Seedance and Nano Banana costs, checked side by side

By SeedRouter · Published 2026-10-05 · Updated 2026-10-05

fal API pricing is set per model: each endpoint page states its own rate per second of video or per image, sometimes as a rough figure backed by a token formula, sometimes with multipliers for higher resolutions. This guide collects those per-model rates for [Seedance 2.5](https://seedrouter.ai/models/seedance-2-5), [Seedance 2.0](https://seedrouter.ai/models/seedance-2-0) and Nano Banana from fal's model pages on October 5, 2026, and lines them up with SeedRouter's rates for the same tiers.

For fal's free credits, credit expiry and the wider list of platforms, see our [fal.ai alternatives guide](https://seedrouter.ai/blog/fal-ai-alternatives). SeedRouter is our product, so read its column knowing the source, and check both pages again before you commit a budget.

Third-party prices in this guide were checked on October 5, 2026.

## How is fal API pricing set?

Each fal model API has its own price, shown on its model page. Video models are billed per second of output, and fal adds a token formula behind several per-second figures, which is why some of them are marked "roughly". Image models are billed per image, and fal states multipliers for larger sizes: [Nano Banana Pro](https://seedrouter.ai/models/nano-banana-pro) 4K is charged at double the standard rate, and [Nano Banana 2](https://seedrouter.ai/models/nano-banana-2) at 1.5 times for 2K and 2 times for 4K.

Three billing rules from fal's FAQ also shape a bill: cold starts are not charged on its model APIs, server errors are never charged, and a client-side error such as an invalid input can still be charged if GPU time was already used.

## What does fal charge for Seedance?

fal's figures are the per-second rates on each Seedance model page for text and image input, marked "\~" where fal itself says "roughly". SeedRouter's are per-second prices for 16:9 video on its model pages.

| Model and resolution    | fal API     | SeedRouter | Difference |
| ----------------------- | ----------- | ---------- | ---------- |
| Seedance 2.5, 480p      | \~$0.2205/s | $0.117/s   | −47%       |
| Seedance 2.5, 720p      | \~$0.4730/s | $0.263/s   | −44%       |
| Seedance 2.5, 1080p     | \~$1.164/s  | $0.647/s   | −44%       |
| Seedance 2.0, 720p      | $0.3034/s   | $0.172/s   | −43%       |
| Seedance 2.0, 1080p     | $0.682/s    | $0.426/s   | −38%       |
| Seedance 2.0 Fast, 720p | $0.2419/s   | $0.118/s   | −51%       |
| Seedance 2.0 Mini, 480p | \~$0.0721/s | $0.018/s   | −75%       |
| Seedance 2.0 Mini, 720p | \~$0.1547/s | $0.039/s   | −75%       |

One minute of 1080p Seedance 2.5 video is about $69.84 at fal's rate. On SeedRouter it costs about 44% less.

## What does fal charge for Nano Banana?

| Model and size           | fal API | SeedRouter | Difference |
| ------------------------ | ------- | ---------- | ---------- |
| Nano Banana Pro, 1K / 2K | $0.15   | $0.03      | −80%       |
| Nano Banana Pro, 4K      | $0.30   | $0.03      | −90%       |
| Nano Banana 2, 1K        | $0.08   | $0.03      | −63%       |
| Nano Banana 2, 2K        | $0.12   | $0.03      | −75%       |
| Nano Banana 2, 4K        | $0.16   | $0.03      | −81%       |

fal's 2K and 4K figures apply its stated multipliers to the listed $0.15 and $0.08 base prices. SeedRouter's standard model IDs for both models charge one flat price per delivered image at any size. A batch of 100 Nano Banana 2 images at 2K is $12 on fal. On SeedRouter the same batch costs a quarter of that.

## Frequently asked questions

### Does fal AI cost money?

Yes. fal model APIs are paid per output from prepaid credits, at the rate on each model page. Its FAQ says purchased credits expire 365 days after purchase; SeedRouter credits never expire.

### What does fal AI do?

fal runs generative media models, such as image, video and audio models, behind its own APIs, and also rents GPUs for custom deployments. Each model has its own endpoint and price.

### How do I use the fal AI API?

You call a model's endpoint with an API key from a server, never from client-side code, and fal bills the request at that model's rate. On SeedRouter, media requests also run as tasks that you poll until they finish; see the [task docs](https://seedrouter.ai/docs/api/tasks).

### Who does fal AI compete with?

Other media APIs that sell the same models, such as Replicate, Kie.ai, WaveSpeedAI and SeedRouter. Our [fal.ai alternatives guide](https://seedrouter.ai/blog/fal-ai-alternatives) compares them on free credits and credit expiry.

### Is fal AI trustable?

fal publishes per-model prices, API documentation and a billing FAQ that states when requests are and are not charged. As with any API, test it with your own traffic before relying on it in production.

## Reading fal API pricing against SeedRouter

fal API pricing is clear per model, and fal hosts a larger catalog plus GPU rental. For the Seedance and Nano Banana tiers both platforms list, SeedRouter's October 5, 2026 rates were 38% to 90% lower. If those models are most of your fal bill, compare the two model pages and move the tiers with the largest gap first.