Reference

Models, modes & credits

Pick the mode by what the shot needs, then pick the engine by the look you want. Gen-2 is the free workhorse; Gen-3 and Real Motion 3.0 are for when the shot has to sell.

Duration

Start at 4s

Resolution

Draft at 720P

Quality tier

Standard first

Every generation mode

ModeCreditsLengthResolutionAudioBest for
Lip-Motion Gen-20Up to 2 min720pUpload up to 2 minLong talking takes, class drafts
Lip-Motion Gen-3~30~10sHigher, upscalableNative, natural syncHero close-ups, final delivery
Image → VideoVariesFrom 4s720P and upNone unless addedAnimating a still, b-roll, loops
Text → VideoVariesMulti-scene720P and upAdd from Voice tabScripted scenes with no source photo
Act-Motion transferVariesReference length720P and upNoneDance, gesture, performance
Real Motion 3.0Higher tier~5s1080p–2KGenerated in the same passFinished cinematic shots with sound

No voice input means a silent video — plan the audio before you render. Upscaling is applied at export.

Native & guest engines

ModelTypeGood for
Real Motion 3.0Native video + audioCinematic shots with synced sound, depth and pose control
Sora 2 / Sora ImageVideo / ImageHigh-realism cinematic shots
Veo 3.1VideoLonger, more coherent motion sequences
RunwayVideoStylised and experimental motion
Nano Banana / ProImagePhotoreal detail, accurate lighting and materials
Flux / Flux 2ImageFast, versatile stills
Z-ImageImageAlternative stylistic range

Credit costs, free allowances, plan pricing and model availability change over time, and some engines sit behind specific plan tiers. Check the dashboard before building a lesson or a course budget around one of them.