Video engine
Veo 3.1 — Google's flagship, with sound
Veo 3.1 is Google's flagship video model: 4, 6 or 8 second clips at 720p or 1080p, with dialogue, ambience and effects generated together with the picture. It reads up to three reference pictures for identity and takes a first and a last frame, a seed and a negative prompt. 106 credits / second with sound — the same at 720p and 1080p — on the Creator plan and up.
50 free credits · no card · adults only, by policy
- Price
- 106 credits / second with sound, at 720p or 1080p alike · silent costs less
- Length
- 4, 6 or 8 seconds
- Resolution
- 720p · 1080p (default)
- Formats
- 16:9 · 9:16
- Sound
- Native — dialogue, ambience and effects · Soundtrack toggle
- Inputs
- Text · start frame · end frame · up to 3 reference pictures · negative prompt · seed
- Access
- Creator plan and up ($49/month)
- Content
- Moderated upstream — SFW work only
What Veo 3.1 is for
Veo 3.1 is the engine for clips that will be watched with the sound on. The soundtrack is generated with the picture — room tone, footsteps, a door, a line of dialogue — so an ad, a short or a character beat arrives with its sound instead of waiting for a sound pass.
Eight seconds is its ceiling, so think in shots rather than scenes: one beat, finished. Longer pieces are a Film mode job — storyboard the scenes and render each one as its own Veo shot.
How to prompt it
- →Write the sound as precisely as the picture: "rain on a tin roof, a kettle starting to whistle" beats "ambient sounds".
- →Keep dialogue short, in quotes and attributed — @Theo says "Wait for me." One or two lines fill an 8-second clip.
- →@mention your cast and elements: up to three of their pictures ride as references, anything beyond rides as text.
- →Put what should not appear in the Avoid field — the negative prompt — instead of negating inside the prompt.
- →Need an exact opening or closing image? Set a start or end frame and describe only the motion between them.
Where it sits in the Veo family
Veo 3.1 Fast renders the same 4 to 8 second clips with the same controls — three reference pictures, frames, seed, negative prompt — quicker and for far less, from the Hobby plan. Veo 3.1 Lite keeps the sound and the frames on every account but reads no reference pictures. Draft on Fast; render on 3.1 when the shot matters.
Starting points
Every tile opens the studio with the prompt already written.
Frequently asked questions
Does Veo 3.1 generate audio?
Yes. With the Soundtrack toggle on, dialogue, ambience and sound effects are generated together with the picture. Turn it off for a silent clip, which costs less per second.
How long can a Veo 3.1 clip be?
4, 6 or 8 seconds. Veo renders in two-second steps, and the studio snaps any other length onto that grid before pricing.
How much does a Veo 3.1 clip cost?
106 credits / second with sound, the same at 720p and 1080p: 424 credits for 4 seconds, 636 credits for 6 and 848 credits for 8. A failed render refunds automatically.
Is there a Veo 3.1 API?
Yes. Send model "veo-3-1" to POST /v1/videos/generations with your eroq key — the same engine, controls and credit balance as the studio, documented in the video API reference.
Can Veo 3.1 render mature content?
No. Veo is moderated upstream and serves SFW work only. Mature fiction between adult characters renders on Motion One, eroq's own uncensored engine.
Keep exploring