Veo 3.1

Short, photoreal, with sound

Veo 3.1 is Google’s video model, and in ClipSpeed it is the specialist for short, polished shots. Its light and texture are its signature: liquid, glass, skin and fabric look like they were shot on a real lens. Clips are 4, 6 or 8 seconds, at 720p or 1080p, in 9:16 or 16:9.

Native audio

Choose a quality marked “audio” and Veo generates sound together with the picture — the sizzle in the pan, rain on the street, a saxophone under a bridge. Because the sound is generated with the motion, it lines up with what you see. Prompts that name the sound (“include the sizzle”) get better results. If you will put music or voiceover on top anyway, pick a silent quality and spend fewer credits.

Where it fits in an edit

Veo is ideal for the premium inserts in a short or an ad: the macro pour, the hero product moment, the four-second hook before the talking starts. Pair it with UGC-style clips from Seedance for the creator look, or with Kling 3.0 for big physical action. For long continuous takes, Seedance 2.5 runs up to 30 seconds.

Credits

Veo 3.1 credits depend on length and audio, not resolution: 720p and 1080p cost the same, so 1080p is usually the right choice. Every option shows its credits before you confirm. Read the native audio guide for how Veo compares with other models on sound.

Prompting Veo

Veo rewards specific, physical description. Describe materials and light (“condensation on cold glass”, “warm lamp light on brass”), keep to one moment per clip, and if you want sound, name it: “include the sizzle”, “rain on the window”, “a crowd cheering in the distance”. For a hook, put the most striking frame first — Veo’s clips are short, so the first second has to work on its own.