AI Models · Google DeepMind
Veo 3.1 — cinema you can hear.
Veo 3.1 is Google DeepMind's flagship video model, and the only family in King AI's video roster that generates dialogue, ambience and sound effects in the same pass as the picture. It runs at 1080p or 4K, in clips of four, six or eight seconds, with no waitlist and no separate Google account.
Specs
What Veo 3.1 actually does.
Every figure below is read from the app's own model catalog at build time, so it is the number you will be charged — not a rounded marketing estimate.
| Mode | Engine | Max resolution | Clip length | Audio | Credits |
|---|---|---|---|---|---|
| Text to video | Veo 3.1 | 1080p | 4s·6s·8s | Yes | 4 per second (~$1.60) |
| Image to video | Veo 3.1 | 1080p | 4s·6s·8s | Yes | 4 per second (~$1.60) |
| Text to video, 4K | Veo 3.1 4K | 4K | 4s·6s·8s | Yes | 6 per second (~$2.40) |
| Image to video, 4K | Veo 3.1 4K | 4K | 4s·6s·8s | Yes | 6 per second (~$2.40) |
| Text to video, fast | Veo 3.1 Fast | 1080p | 4s·6s·8s | Yes | 1.5 per second (~$0.60) |
| First & last frame | Veo 3.1 First-Last Frame | 1080p | 4s·6s·8s | Yes | 4 per second (~$1.60) |
| Reference to video | Veo 3.1 Reference | — | 8s | Yes | 4 per second (~$1.60) |
| Extend a clip | Veo 3.1 Extend | — | 7s | Yes | 4 per second (~$1.60) |
Strengths
What Veo 3.1 does best.
Google's flagship video model — cinematic realism with native audio.
Sound in the same pass
Dialogue, room tone and effects are generated with the frames, so there is no separate audio job to line up afterwards.
A real 4K tier
Veo is one of three families here — with Seedance and Kling — that generates 4K natively rather than upscaling to it afterwards.
Six ways in
Text, image, first-and-last frame, reference, extend, and a cheap fast tier — the widest mode coverage of any family in the app.
A draft tier that is genuinely cheap
Veo 3.1 Fast bills at a fraction of the flagship rate, so exploration and delivery use the same model at two prices.
Limits
Where Veo is the wrong pick.
No model wins every shot. These are the cases where something else in the app is the better spend.
Eight seconds is the ceiling. Veo generates 4, 6 or 8 second clips and nothing longer, so a single take of 12 or 15 seconds has to come from another model — or from Veo plus the extend mode, which adds seven seconds at a time.
It is the second most expensive way to fill a second of 1080p in this line-up at 4 credits, and the 4K tier adds 2 credits a second on top. For a shot you are still exploring, generate on Veo 3.1 Fast at 1.5 credits a second and re-run the keeper on the flagship.
Reference-to-video is fixed at eight seconds, so a reference-driven shot cannot be shortened to trim cost.
Head to head
Veo 3.1 against the alternatives.
The same six columns for the models you would realistically pick instead, on the same job.
| Model | Max resolution | Clip length | Audio | Credits per second | Shortest clip |
|---|---|---|---|---|---|
| Veo 3.1 | 1080p | 4s·6s·8s | Yes | 4 (~$1.60) | 4s · 16 credits (~$6.40) |
| Kling V3 Pro | 1080p | 3–15s | Yes | 1.68 (~$0.67) | 3s · 5.04 credits (~$2.02) |
| Seedance 2.0 | 1080p | 4–15s | Yes | 6.804 (~$2.72) | 4s · 27.22 credits (~$10.89) |
| MiniMax H3 | 768p | 5s·10s·15s | Yes | 0.6 (~$0.24) | 5s · 3 credits (~$1.20) |
Best for
Where to use Veo 3.1.
Every card links to the King AI tool that gets you there fastest.
Dialogue scenes
Characters that speak, with the audio generated alongside the picture.
Delivery in 4K
Hero shots that go out at native 4K rather than an upscale.
Premium ad spots
Brand work where the finish has to read as graded footage.
Photo animation
High-fidelity image-to-video that keeps the subject on model.
How it works
Getting the most out of Veo 3.1.
Pick the tier before the prompt
Veo 3.1 Fast for exploration, Veo 3.1 for delivery, Veo 3.1 4K when the shot is going out large. Same prompt box, three prices.
Write the sound in
Because Veo generates audio with the picture, naming the sound — a line of dialogue, rain on a window, a room's echo — changes the result. Silent prompts get generic ambience.
Choose 4, 6 or 8 seconds
The duration is the price. Eight seconds costs twice four, so start short and lengthen only the take you keep.
Extend, don't regenerate
Veo's extend mode adds 7 seconds to an existing clip and inherits its look, which is cheaper than re-rolling a longer prompt.
Prompts
Three prompts that play to Veo 3.1's strengths.
Written for this engine, not swapped between pages. Each one opens the studio with Veo 3.1 selected and the prompt already in the box — the cost shown is what one clip at the shortest length it offers will take off your balance.
A line of dialogue
“A rain-soaked detective leans into a payphone under a flickering sign and says, ‘I found the boat.’ Neon bleeding across wet asphalt, thunder rolling behind the line, handheld 35mm, no music.”
4s · 16 credits (~$6.40)
Generate on Veo 3.1 →Sound that is the subject
“A glassblower turns a molten orb in a dark workshop while the furnace roars and the glass ticks as it cools. Warm rim light, slow push-in, shallow depth of field, breath audible between turns.”
4s · 16 credits (~$6.40)
Generate on Veo 3.1 →An establishing shot in 4K
“Dawn over terraced rice fields, mist unpeeling from the water, a farmer’s footsteps and distant birdsong. Slow crane up to a wide vista, natural light, no score, no camera shake.”
4s · 16 credits (~$6.40)
Generate on Veo 3.1 →Proof
Same prompt, this engine.
One prompt was sent to eight engines from the same starting frame. This is what Veo 3.1 did with it. The prompt, word for word: “A paper boat the size of a ship sails through a flooded cathedral at dawn, candle flames guttering as it passes, water lapping the stone, slow lateral dolly, 35mm, natural light, the sound of water and wind, continuous motion, no cuts, no new objects”
Veo 3.1, run on the prompt above at 1080p · 8s · 32 credits (~$12.80). Tap or scroll and the still becomes the clip.
Keep exploring
Related models & guides.
Same-prompt comparisons
Every flagship pair on one photo and one prompt, with the cost of each take.
ExploreVeo 3.1 cost per video
What one clip costs here, next to the maker's own plans.
ExploreImage to video with Veo 3.1
Animate a photo with Veo 3.1: the specs, the cost and what the model does with a still.
ExploreKling V3 Pro
The motion king — fluid, physical movement at unbeatable value.
ExploreSeedance 2.0
Multi-shot storytelling, with a native 4K tier and a cheap draft tier.
ExploreWAN 3.0
The longest single take in the app — 30 seconds, billed by the second.
ExploreVeo 3.1 vs Kling — full comparison
See how the flagships stack up on real prompts.
ExploreFAQ
Veo 3.1, answered.
How can I use Veo 3.1 without a waitlist?
Download King AI or open the Web Studio. Veo 3.1 is in the standard model list on iPhone, iPad, Mac and the web, billed from the same credit balance as every other model. There is no Google account, no waitlist and no separate Veo subscription.
What resolution and clip length does Veo 3.1 support?
Veo 3.1 outputs 1080p and Veo 3.1 4K outputs 4K. Both generate clips of 4, 6 or 8 seconds. 8 seconds is the hard ceiling for a single Veo generation; longer scenes are built by extending, which adds 7 seconds at a time.
How much does a Veo 3.1 video cost in King AI?
King AI bills per second. Veo 3.1 costs 4 credits per second, so a 4-second clip is 16 credits and an 8-second clip is 32. Veo 3.1 4K costs 6 credits per second. Veo 3.1 Fast costs 1.5 credits per second, which makes a 4-second draft 6 credits. In most countries, your first video is on us: new verified accounts get one five-second video on our lowest-cost video model, once. A four-second Veo 3.1 run costs more than the first video on us, so it needs credits from a plan or credit pack.
Does Veo 3.1 generate audio?
Yes. Veo generates dialogue, ambience and sound effects in the same pass as the picture, which is why naming the sound in your prompt changes the result. Most models in the catalog produce a silent clip that needs a separate audio job.
Is Veo 3.1 better than Kling V3 Pro or Seedance 2.0?
It depends on the shot. Veo has a native 4K tier and the strongest audio, but caps at 8 seconds. Kling V3 Pro reaches 15 seconds and costs 1.68 credits a second against Veo's 4. Seedance 2.0 reaches 15 seconds and holds a multi-shot story together. In King AI you can run the same prompt on all three and keep the one that wins.
Get started
Try Veo 3.1 today.
In most countries, your first video is on us: new verified accounts get one five-second video on our lowest-cost video model, once. A four-second Veo 3.1 run costs more than the first video on us, so it needs credits from a plan or credit pack.