Models & prices
One credit is $0.01 of what we pay the model vendor. Credits are spent only when a render succeeds — a failure costs nothing, and polling a job is free.
These numbers are computed by the same code that bills the job, so the table and the invoice cannot disagree. For an exact figure on a specific call, ask your agent for estimate_cost.
Image
priced for one 1:1 imagePriced per render. A wider aspect ratio on a megapixel-priced model costs proportionally more.
| Model | Credits | Aspect ratios | Notes |
|---|---|---|---|
Seedream fal-ai/fal-ai/bytedance/seedream/v4.5/text-to-image | 4 cr | 1:1, 16:9, 9:16, 4:5 | |
Flux fal-ai/fal-ai/flux-2-pro | 4 cr | 1:1, 16:9, 9:16, 4:5 | |
Flux Max fal-ai/fal-ai/flux-2-max | 8 cr | 1:1, 16:9, 9:16, 4:5 | |
Nano Banana fal-ai/fal-ai/nano-banana-2 | 8 cr | 1:1, 16:9, 9:16, 4:5 | |
Nano Banana Pro fal-ai/fal-ai/nano-banana-pro | 15 cr | 1:1, 16:9, 9:16, 4:5 | |
GPT Image 2 openai/gpt-image-2 | 20 cr | 1:1, 16:9, 9:16, 4:5 |
Video
priced for 6 secondsPriced by duration, so a longer clip costs more than the figure quoted here.
| Model | Credits | Aspect ratios | Notes |
|---|---|---|---|
Fast, cheapest fal-ai/pixverse/v4.5/text-to-video | 10 cr | 16:9, 9:16, 1:1 | up to 8s |
Higher quality, cinematic fal-ai/bytedance/seedance/v1/lite/text-to-video | 13 cr | 16:9, 9:16, 1:1 | up to 10s |
Veo 3.1 Fast (with audio) kie/veo3_fast | 40 cr | 16:9, 9:16 | up to 8s |
Veo 3.1 (with audio) kie/veo3 | 200 cr | 16:9, 9:16 | up to 8s |
Speech
priced for 1,000 charactersPriced by the length of the text you send, not the length of the audio produced.
| Model | Credits | Aspect ratios | Notes |
|---|---|---|---|
Kokoro TTS fal-ai/fal-ai/kokoro | 2 cr | — |
Not sure which to pick?
You do not have to. Leave the model out and SuperGen routes the request — ask for prefer="cheap" and it takes the lowest real credit cost that can do the job, prefer="quality" for the top of the range. Plans and included credits are on Pricing.