A chorus, the candles and a cheer in one Veo 3.1 clip
Veo 3.1 is the full model of Google's Veo line and the top of the three Veo options in our picker. It turns a photo of your people into a 4, 6 or 8 second clip and makes the sound in the same pass, voices included: a family finishing Happy Birthday, one short toast, a room that claps. Veo 3.1 Fast and Lite take the same settings and cost less per second, and for plenty of clips they are enough.
- Made by Google
- from 150 credits
- 4-8 s
- With or without sound
Updated 29.09.2026
Two clips on the full Veo 3.1, one on Fast
Prompt
The family gathered around the birthday cake softly sings the last line of Happy Birthday, then Mum leans in and blows out the candles, everyone cheers and claps, candle smoke curls up, warm cosy kitchen light, slow cinematic push-in
Prompt
The newlyweds on the seaside terrace clink their glasses and share a quiet laugh, the guests behind them cheer, sea waves roll in at sunset, a gentle breeze lifts the bride's veil, cinematic slow dolly, sounds of waves, glasses and cheering
Prompt
Friends on the rooftop raise their glasses in a toast as the city lights come on at dusk, one says cheers and they all laugh, string lights sway, smooth slow orbit camera, sounds of the city and clinking glasses
Veo 3.1, Fast or Lite: where the extra credits go
On paper the three are the same tool. They take the same 4, 6 or 8 seconds, the same two sizes, the same sound switch, wide or tall frames. Fast is the quicker, cheaper version and Lite the budget one. The full Veo 3.1 costs the most per second and carries the Cinema quality tag.
Our rule for spending the difference: close faces, hard light, and a clip that will be watched big or many times. The candles in a dark kitchen and the wedding toast into a setting sun were both made on the full model at 1080p.
The rooftop toast was made on Fast at 720p, and on a phone it holds up: three friends, glasses meeting more than once, a skyline going gold behind them. The frame is wide, the light is soft and the audience is a group chat. Lite is where we rehearse: run the prompt there to check the timing, then make the keeper on the full model. One more point in the full model's favour: its 1080p costs no more per second than 720p at today's prices, so there is no reason to go smaller.
Writing voices for Veo 3.1
Veo 3.1 can make people in the clip sing and speak, and it follows plain instructions. Say who makes the sound and when: "the family softly sings the last line of Happy Birthday, then Mum leans in and blows out the candles, everyone cheers". That is close to the prompt under the candles clip, and the clip plays it in that order.
Spoken words go in quotes, one short sentence per clip. Eight seconds is the most you get, and a line that takes five of them leaves no room for the reaction. Name the speaker, "Dad turns to the camera and says", so Veo knows whose line it is.
Two limits to know before you spend. The voices are Veo's own invention: it has never heard Grandma and cannot copy her. Our examples are in English; if your line is in Hebrew, Arabic or Russian, keep it to a few words and watch the result before sending. When it has to be your text in a voice you chose, switch to a talking card, which reads it with ElevenLabs voices, Hebrew included. A full song, verses and chorus, is a job for Lyria.
Second by second: the three clips above
The candles clip runs 6 seconds. For the first two, Mum laughs with her hands at her face while the song ends. At about two and a half seconds she leans in and blows. By the fourth second the smoke is curling up into the lamp light, and the last two seconds are applause and smiles. That is one breath and one cheer, with a second of room on either side.
The wedding clip runs 4 seconds and spends only the first second and a half on the toast. After that the glasses come down and everyone looks at each other while the waves roll in behind them. Four seconds was enough, and it is the cheapest length there is.
The rooftop toast on Fast runs 6 seconds: a clink, laughter, another clink, the sun sinking behind the towers. Count the beats of your own moment the same way before you choose a length, and you will rarely pay for seconds nobody needed.
What the full Veo 3.1 is for
People can sing and say a line
Veo 3.1 puts voices into the clip along with the picture. In the candles clip above the family sings the end of Happy Birthday, Mum laughs, blows, and the kitchen applauds. A short line in quotes works too: "surprise", "cheers", "happy birthday, Mum".
Hard light, close faces
Candlelight in a dark kitchen, a low sun right behind the bride: this is the light we hand to the full model, tagged Cinema quality in our picker. Both clips above that sit close to faces in difficult light were made with it at 1080p.
The sound lands on the action
Picture and audio come out of one generation, so they agree. The smoke rises after the breath, the applause starts as the hands meet, the clink sounds when the glasses touch. There is nothing to line up afterwards, and no stock tune pasted under the moment.
Four seconds is often enough
The full model costs more per second than anything else in our video picker, so spend seconds with care. The seaside wedding above is 4 seconds long: the glasses meet in the first second and a half, then everyone just smiles while the sea does the rest.
Prompts for Veo 3.1 with voices in them
Dad at the garden grill turns to the camera with the tongs raised and says "Happy birthday, Nadia, dinner is on me", the kids behind him cheer, smoke drifts through the evening sun, sounds of sizzling, birds and laughter
Three cousins squeezed onto a sofa sing the last line of Happy Birthday in terrible harmony, the youngest goes for a high note and they all collapse laughing, lamp light, handheld camera, sounds of singing and giggling
Two glasses meet over a candlelit table on a balcony above the harbour, the couple smile at each other, boat lights sway on the water below, slow dolly, sounds of lapping water, a distant ship's horn and the clink
Setting up a Veo 3.1 clip
- 1
Attach the first frame
Press the button on this page, or switch the composer to Standard and choose Video. Attach one picture: a photo of your people or a card you made here with all of them in it, as JPG, PNG or WebP up to 10 MB.
- 2
Choose Veo 3.1, not Fast, Lite or Auto
Open Model and pick Veo 3.1 itself; Auto would give you MiniMax H3. Set 4, 6 or 8 seconds and 1080p, which on the full model costs the same per second as 720p today. Leave sound on if anyone should sing or speak.
- 3
Write the action, then the sound, then the words
Describe what moves, then what we hear, then any spoken line in quotes. The price sits on the button until you press it, and a failed clip returns its credits. Send the result by email, SMS or a popular messenger, now or on the day.
Veo 3.1 beside Fast and Lite, priced today
The three rows share every setting, so the price column decides: the full Veo 3.1 for close faces and hard light, Fast for the group chat, Lite for a rehearsal. For more than 8 seconds or a square frame, open Kling 3.0 instead.
| Model | Strong at | Length | Quality | Sound | Price |
|---|---|---|---|---|---|
| Veo 3.1Google | Cinema quality | 4, 6, 8 s | 720p, 1080p | With or without sound | from 1,000 credits · 500 cr a second |
| Veo 3.1 FastGoogle | Fast | 4, 6, 8 s | 720p, 1080p | With or without sound | from 500 credits · 188 cr a second |
| Veo 3.1 LiteGoogle | Budget | 4, 6, 8 s | 720p, 1080p | With or without sound | from 150 credits · 63 cr a second |
Prices are in credits and are always shown on the button before you press it. Pricing
Keep exploring
Veo is Google's video model, here in three versions: Veo 3.1, Veo 3.1 Fast and Veo 3.1 Lite. Each turns…
OpenKling 3.0Kling 3.0 Pro is Kuaishou's video model, and of the three Kling models we run it is the only one…
OpenWanWan 3.0 Prime is Alibaba's video model, and on Wish Generator it is the one we point to for small,…
OpenAI video greetingAn AI video greeting starts from a photo of your people and ends as a short clip where they move…
OpenQuestions and answers
Price and finish, not settings. Both take 4, 6 or 8 seconds, 720p or 1080p and a sound switch. Fast is the quicker, cheaper version; the full Veo 3.1 costs more per second and is the one we use for close faces in difficult light, like the candles and the wedding above. For a wide group shot watched on a phone, Fast is usually enough.
Yes. Veo 3.1 generates voices along with the picture: singing, a short spoken line, cheering. Write who speaks and put the words in quotes, one sentence at most, because a clip is 8 seconds at the longest. The candles clip on this page ends with the family singing Happy Birthday and cheering, all from one generation.
No. Veo 3.1 invents the voices, since it has never heard your family. If the words must be said in a particular voice, make a talking card instead: a photo of the person speaks your text in a voice you pick from our catalogue, for up to 15 seconds. For a sung greeting with verses and a chorus, make a song with Lyria.
Because most greetings do not need it. Auto picks MiniMax H3, which costs far less per second, always has sound and animated every video scene in our catalogue. Veo 3.1 is the model to choose by hand when voices, close faces or difficult light carry the clip: open Model in Standard mode and pick it.
1080p. On the full Veo 3.1 the two sizes cost the same per second at today's prices, so the bigger frame comes at no extra cost, and the same holds for Veo 3.1 Fast. Only Veo 3.1 Lite charges more for 1080p, and there 720p is a sensible saving for a draft.
Hear them sing before the candles go out
Attach a photo of your people, choose the full Veo 3.1, and have the clip arrive on the morning of the day.
Make it with Veo 3.1
