Veo 3.1
Video model · Google ·
Veo 3.1 is a video model by Google. In Endless it costs from 13 credits per video and is available on every plan.
Veo 3.1 is Google's text and image-to-video model for cinematic 4, 6 or 8 second shots at up to 4K, with optional native audio.
- Price
- From 13 credits per video
- Availability
- Available on every plan
- Duration
- 4, 6 or 8 s
- Resolution
- 720p, 1080p or 4K
- Inputs
- Text, image, first and last frame
- References
- Up to 3 images
- Audio
- Optional, native
About Veo 3.1
Veo 3.1 is Google DeepMind's video model for short, cinematic shots. It generates clips of 4, 6 or 8 seconds in 16:9 or 9:16 at 720p, 1080p or 4K. Turn audio on and it creates dialogue, sound effects and ambient noise together with the picture; spoken dialogue works best in English. Every clip carries a SynthID watermark that marks it as AI-generated.
You can describe a shot from scratch, animate a starting image, or give a first and last frame and let Veo build the transition between them. Up to three reference images keep a character, object or style consistent, which helps when the same mascot or product appears across a series of clips. Teams use it for product hero shots and vertical cuts for social feeds.
Best for
- Short cinematic inserts and establishing shots
- Vertical 9:16 clips for social feeds
- Transitions between a first and last frame you choose
- Keeping a product or character consistent with reference images
How to prompt Veo 3.1
How Veo 3.1 reads a prompt
Veo 3.1 makes one shot, or a short run of shots, lasting 4, 6 or 8 seconds. Google's prompt guide breaks a prompt into elements: subject, action and style first, then camera position and motion, composition, focus and lens effects, and ambiance. Google asks for descriptive, specific language, and its examples refine a short prompt by adding concrete detail about the face, the light, the lens and the setting, not words about quality. Because the clip is short, write for one moment and describe it fully, rather than a story that needs a minute to tell.
A structure that works
Google Cloud's prompting guide proposes a five-part formula:
[Cinematography] + [Subject] + [Action] + [Context] + [Style and ambiance]
For example: "Low-angle tracking shot, a cyclist in a yellow rain jacket, pedaling hard up a cobbled hill, a narrow old-town street after rain with wet stones reflecting neon signs, cool blue tones, shot on 35mm film." Lead with the camera when the shot depends on it. If audio is on, add the sound at the end as its own sentence.
For more than one shot in a single clip, Google describes timestamp prompting: split the clip into segments such as [00:00-00:02] and [00:02-00:04], and give each its own shot, action and sound. Keep the character description identical in every segment so the person stays the same.
Camera and motion vocabulary
Veo understands standard film terms, and Google's guide calls cinematography the most powerful part of the prompt for tone and emotion.
- Movement: dolly shot, tracking shot, crane shot, aerial view, slow pan, POV shot.
- Composition: wide shot, close-up, extreme close-up, low angle, two-shot, top-down.
- Lens and focus: shallow depth of field, deep focus, soft focus, macro lens, wide-angle lens.
- Light and color: golden hour, natural light, warm tones, cool blue tones, neon glow.
Pick one main movement per clip and say what it reveals ("the crane rises to reveal the canyon"). An 8-second clip rarely has room for more than one or two moves.
Audio and dialogue
Audio is off by default in Endless. Turn on Generate audio and Veo creates dialogue, sound effects and ambience together with the picture. Google's guidance for each:
- Dialogue in quotation marks: A woman says, "We have to leave now."
- Sound effects described explicitly: SFX: thunder cracks in the distance.
- Ambience as its own line: Ambient noise: the quiet hum of a starship bridge.
Keep spoken lines short. fal's guide notes that long speeches in an 8-second clip get cut off or rushed. Google says English is fully supported and other languages have not been evaluated, so dialogue in other languages may vary. If you want silence or no music, say so. Google also notes that Veo sometimes blocks a generation because of safety filters or a problem processing the audio; if that happens, reword the dialogue or try with audio off.
Images, first and last frames, and references
There are three ways to guide Veo with images:
- Start from an image. Attach one image and turn on First frame, and Veo animates from it. Google suggests picking the image closest to the first scene you want. In the prompt, describe the motion and sound, not what is already in the picture.
- First and last frame. Add a last frame as well and describe the transition between the two, such as a camera arc from a close-up to a wide view.
- Reference images. Attach up to three images of a person, character, product or style, and Veo keeps their appearance. Say what each one is ("using the provided images for the detective, the woman and the office").
Every Veo video carries a SynthID watermark that marks it as AI-generated.
Settings that matter
- Mode: Standard for the highest fidelity, Fast as the balanced default, Lite as the lightest tier, up to 1080p.
- Duration: 4, 6 or 8 seconds. In Google's API, 1080p, 4K and reference-image generations run at 8 seconds.
- Resolution: 720p, 1080p or 4K. Google notes that higher resolution takes longer to render.
- Aspect ratio: 16:9 or 9:16.
- Generate audio: off by default.
- Image, First frame, Last frame and reference images: as described above.
Common failure modes and fixes
- The result looks generic. The prompt is short and vague. Add camera, lens, light and a clear style.
- An exclusion is ignored. Describe what you want instead, as Google suggests: "a desolate landscape with no buildings or roads" rather than "no man-made structures".
- Dialogue is rushed or cut off. Shorten the line or choose a longer duration.
- A character changes between clips. Reuse the exact same description, or attach reference images.
- Too much happens in one clip. An 8-second clip has room for a few actions. Split a longer story into several clips and edit them together.
Questions about Veo 3.1
- What is Veo 3.1?
- Veo 3.1 is a video model by Google.
- Is Veo 3.1 free?
- You can try Veo 3.1 for free. It is on every Endless plan, including Free, which gives you 100 credits when you create a workspace and 50 credits every week. Each run spends credits from that balance.
- How much does Veo 3.1 cost?
- In Endless it costs from 13 credits per video and is available on every plan.
- How do I use Veo 3.1?
- Open Video in your Endless workspace, pick Veo 3.1 in the model picker and write a prompt or add a starting image. You can also ask for it by name in chat or use it on a canvas.
- What are the alternatives to Veo 3.1?
- Similar models you can use in Endless include Seedance 2.5, Seedance 2.0, and FLUX 3. Each one has its own page with its price and plan.
- Can Veo 3.1 generate sound?
- Yes. With audio on, Veo 3.1 generates dialogue, sound effects and ambient noise together with the video. Audio is optional, so you can also render silent clips.
- How long are Veo 3.1 videos?
- Each generation is 4, 6 or 8 seconds long. For a longer piece, generate several shots and edit them together.
- Does Veo 3.1 support vertical video?
- Yes. It renders in 9:16 for vertical formats as well as 16:9 for landscape.
How to use Veo 3.1 in Endless
- 01
Open Video
In your workspace, open Video and pick Veo 3.1 in the model picker.
- 02
Describe the shot
Write a prompt with the subject, camera and motion, or start from an image you upload or generate.
- 03
Generate and iterate
Generate, then refine in chat or on a canvas. Every clip is saved to your workspace assets.
Example prompts for Veo 3.1
Grouped by use case. Each one opens in Endless with Veo 3.1, ready to edit and generate.
Product shots
Slow dolly in on a frosted glass bottle of cold-pressed orange juice on a marble counter, condensation beading on the glass, morning sun raking from the right, shallow depth of field, clean commercial look. SFX: a soft clink as an ice cube settles.
Macro tracking shot gliding along the bezel of a steel dive watch resting on wet black rock, waves washing over it and draining away, cool teal light, crisp detail on the engraved numerals.
Top-down shot of a wooden cutting board as hands slice a ripe avocado in two and twist the halves apart, bright natural kitchen light, fresh and minimal. Ambient noise: a quiet kitchen, a knife on wood.
Low-angle orbit around a white electric scooter parked on a rooftop at golden hour, the city skyline soft in the background, lens flare as the camera passes the sun, warm cinematic grade.
Close-up of a hand pressing a fountain pen to cream paper and writing one looping signature, ink spreading into the fibers, shallow focus, warm desk lamp. SFX: the scratch of the nib, a page settling.
A pair of noise-cancelling headphones floats and rotates slowly against a deep navy background as thin rings of light pulse outward from the ear cups, studio product lighting, smooth and precise motion.
Establishing shots
Crane up from the wet deck of a fishing boat at dawn to a high view of the whole fleet leaving a fog-covered harbor, soft pastel light, quiet and vast. Ambient noise: gulls, idling engines, water lapping against hulls.
Aerial view sweeping over a red-rock desert canyon at midday, a lone off-road truck kicking up a dust trail along the canyon floor, heat haze, deep blue sky, 70mm epic film look.
Wide shot of a neon-lit ramen stall on a rainy Tokyo backstreet at night, steam rising from the counter, a cook ladling broth, puddles reflecting pink and blue signs, slow pan from left to right. Ambient noise: rain on awnings, a distant train.
Static wide shot of an alpine lake at first light, mist drifting over the water, a single wooden rowboat tied to a dock, the sun cresting the peaks and turning them gold, tranquil and still.
Slow push-in through the open doors of an abandoned 1920s ballroom, dust hanging in shafts of afternoon light, a chandelier swaying slightly, faded wallpaper peeling, haunting and beautiful. Ambient noise: creaking floorboards, faint wind.
Tracking shot from a moving train window across rolling green hills in Scotland under fast-moving clouds, sunlight breaking through and racing across the fields, documentary feel, soft film grain.
Dialogue scenes
Medium two-shot in a cramped recording booth: a young producer in headphones leans toward a nervous singer and says, 'One more take, and this time, mean it.' The singer exhales and nods. Warm practical light, shallow depth of field, the low hum of the room.
Close-up of an elderly fisherman mending a net on a pier at sunset, his apprentice beside him. He looks up and says, 'The sea gives back what you give it.' The apprentice smiles. Ambient noise: waves, creaking boards, distant gulls.
Handheld medium shot in a busy hospital corridor at night: a doctor in scrubs stops a hurrying nurse. Doctor: 'Room six is waking up.' Nurse: 'Then go, I've got this.' Fluorescent light, cool green tones, beeping monitors and footsteps.
Claymation scene of two snails at a tiny café table. The first snail asks, 'Did you order yet?' The second answers, very slowly, 'I'm still deciding.' Soft accordion music, clinking teacups.
Medium shot of a teenage girl on a rooftop at dusk holding a walkie-talkie, city lights coming on behind her. She presses the button and says, 'Tower to base, the lights are working.' A crackly voice answers, 'Copy that, come home for dinner.' Wind, distant traffic.
[00:00-00:04] Over-the-shoulder shot of a barista handing a cup across the counter to a regular customer. [00:04-00:08] Reverse close-up of the customer reading the name on the cup and laughing, 'You spelled it right this time.' Café chatter, espresso machine hiss.
Vertical social clips
Vertical 9:16, handheld POV walking through a crowded night market in Bangkok, steam and smoke rising from grills, vendors calling out, a hand reaching for a skewer of grilled pork, vibrant colors, energetic. Ambient noise: sizzling, chatter, motorbikes.
Vertical 9:16, locked-off close-up of a latte being poured as the milk forms a rosetta, warm overhead café light, shallow focus, slow motion. SFX: a soft pour, the cup set down on its saucer.
Vertical 9:16, a skateboarder drops into an empty concrete pool at sunset, the camera following low and fast from behind, warm backlight, light film grain. SFX: wheels on concrete, a clean landing.
Vertical 9:16, a golden retriever puppy tumbles through autumn leaves in a park, chasing a red ball toward the camera, low eye-level shot, soft afternoon light, playful.
Vertical 9:16, medium close-up of a fitness coach facing the camera in a bright studio. She says, 'Three moves, thirty seconds each, let's go.' Then she drops into the first squat. Upbeat music, clean natural light.
Vertical 9:16, top-down shot of hands arranging travel gear on a linen bed: a passport, a film camera, sunglasses and a paper map, each item placed with a soft tap, bright morning light.
Image-led shots
Start from the attached product photo. The camera slowly orbits the bottle as morning light moves across the label and soft shadows shift on the counter, keeping the label sharp throughout.
Using the attached first and last frames, the camera pulls straight back and rises from a close-up of a sprouting seed to the full garden at sunset, plants growing in time-lapse along the way. Soft wind, birdsong.
Using the provided images for the chef, the kitchen and the copper pan, a medium shot of the chef tossing vegetables in the pan over a high flame, drops of oil catching the light. Sizzling, kitchen clatter.
Animate the attached illustration: the paper lanterns sway gently, the river glimmers and a small boat drifts from left to right, keeping the hand-painted style and colors of the image.
Using the provided image of the mascot, a cartoon version of the mascot surfs a giant wave at sunset and waves to the camera as spray flies past. Bright, saturated, playful. Ocean roar, cheerful ukulele.
Start from the attached portrait. She turns her head slowly toward the window, the curtain lifts in the breeze and light passes across her face, subtle and natural, shallow depth of field.
Related models
Seedance 2.5
by ByteDance
30-second clips in one pass, guided by up to 50 references
From 51 credits per video
Seedance 2.0
by ByteDance
Clips of 4 to 15 seconds steered by image, video and audio references
From 16 credits per video
FLUX 3
by Black Forest Labs
Clips up to 20 seconds with sound, from a prompt, frames or a clip
From 90 credits per video
Compare Veo 3.1
Veo 3.1 next to other models: price, maker and plan side by side.
Give your next idea a place to grow.
Start free, create a project, and bring your brand’s references together. Or book a demo to explore Endless with your team.