Seedance 1 Pro Fast: the still is the contract (5s, 10s, 1080p, 62cr)

Seedance 1 Pro Fast is image-to-video. If the label is wrong, every second is waste.

Seedance 1 Pro Fast is ByteDance's accelerated image-to-video generate with high visual quality. Content type is video. Category is image-to-video only. Requires an image. Audio is off. Durations: 5s and 10s. Max output: 1080p (quality chips 720p and 1080p). Aspects: 21:9, 16:9, 4:3, 1:1, 3:4, 9:16. Display price: 62 credits. The catalog marks it fast. It does not invent the plate.

Image in, five or ten seconds out

This row will not start from a paragraph. requires_image is true. The still you upload is the contract: composition, product, type, the face. Seedance 1 Pro Fast's job is motion on that file for 5s or 10s at up to 1080p.

If you wanted text-to-video, this label is wrong. Every second of the 5s or 10s clip is motion on the wrong brief. The image-to-video term is the job. The tool is AI image-to-video. The ranked list is best image-to-video model.

Fast is a generation speed flag. It is not a reason to skip the still. Acceleration on a bad plate is 62 credits of high-quality wrong.

Silent 1080p is the spec

audio is false. There is no native soundtrack to write into the prompt. If the brief needs dialogue or bed, you generate speech on a TTS row and lay it in the editor. Prompting "voiceover: ..." on this model does not buy you a mix. It buys you a silent 5s or 10s at 1080p with extra words the motion model may ignore or paint as type.

62 credits is the display price. The price matrix is duration-shaped (min 62, max 124). Pick 5s when the still only needs a hold. Pick 10s when the camera move needs the extra seconds. Do not pick 10s because the slider goes there. The still is already the expensive decision; the extra five seconds multiply a miss.

Approve the plate on Seedream, then spend 62

Lock the plate on a stills row first. Seedream 5.0 Pro is the ByteDance stills flagship (text-to-image, 7 credits, no clip). Approve that. Then run Seedance 1 Pro Fast. The ByteDance roster is built as that split. Using Seedance to "find" the look is how you spend 62 credits learning the label was wrong.

The I2V versus T2V fork is which to use. For this row the test is blunt: do you have the still? If no, you are not on Seedance 1 Pro Fast yet.

Match aspect to the plate. A 9:16 still into a 16:9 generate is a crop the model will invent. 1080p will faithfully resolve the crop.

Sixty-two credits of high-quality wrong

Wrong label looks like: text-only prompt on an image-required row; a talking-head brief on a silent model; a 30s need on a 5s/10s row; a still you have not approved. Each of those wastes the 5s or the 10s at 62 credits.

Correct label: approved still, 5s or 10s, 1080p, no audio, ByteDance image-to-video, 62 credits. Generate that. Caption and mix after.

FAQ

Can I run Seedance 1 Pro Fast from text only?

No. The row requires an image. If you only have a paragraph, lock a still first, then come back. Text-to-video is a different category.

Does the fast variant include a soundtrack?

No. audio is false. 5s or 10s, up to 1080p, silent. Add voice on a speech row and mix in the editor.

Why is it 62 credits if I only need a short move?

62 credits is the display price for this image-to-video pass. The 5s option is the short move. The 10s option costs more on the duration matrix. Neither is a stills generate.

What happens if I upload an unapproved plate?

The model treats it as the contract. High visual quality on the wrong label, for every second of 5s or 10s. Approve the still. Then buy motion.