LiveVeo 3.1 is live: reference images, 4K output and scene extension up to ~148 seconds. Try it now →
Veo 3
Google's AI video model.
Veo 3 turns text or an image into a finished video with the sound already in: score, effects and lip-synced dialogue generated in the same pass as the picture. Generate text-to-video and image-to-video in the browser.
What is Veo 3?
Veo 3 is an AI video generation model from Google DeepMind. It was the first mainstream model to generate video and synchronized audio in a single pass — put a line of dialogue in quotes and the character says it with matching lip movement. It reads plain-language direction for camera moves, lighting and tone, and renders physics that hold up to a rewatch.
Released in May 2025 at Google I/O, Veo 3 set the template every video model since has chased, and it remains available on Veo Studio alongside its successor, Veo 3.1. You get a simple prompt box, no downloads, and a free tier to start. Developers can also call it through the Veo API.
Veo 3 features.
Native sound and dialogue
Music, ambient effects and character speech are generated with the picture, synced to the action. Put a line in quotes and your character says it, lip movement included. No separate audio pass, no manual syncing.
Text-to-video and image-to-video
Prompt with words, a photo, or both. Supply an image and Veo 3 animates it into a full shot — your product photo or portrait becomes the opening frame.
Physics that hold up
Sports, dancing, collisions, cloth and water, complex motion renders the way your eye expects, so the clip survives a rewatch.
Speaks the language of film
Dolly-ins, wide establishing shots, neon rim light, shallow depth of field — write direction the way a filmmaker would and Veo 3 executes it. No settings panels, just words.
Veo 3 speaks the language of film.
Camera moves, lighting setups, lens choices, emotional tone, write a Veo prompt the way a director would and the model executes it. No settings panels, just words.
How to use Veo 3.
- 01
Sign in and open the studio
Create a free account and pick Veo 3 from the model selector (Veo 3.1 is the default), no waitlist, no invite code and nothing to download.
- 02
Write a prompt or add an image
Describe the shot, the camera move and any dialogue in quotes. Optionally attach a photo as the starting frame for image-to-video.
- 03
Generate and download
Veo 3 returns a finished clip of 4 to 8 seconds at up to 1080p, with sound. Pick 16:9 or 9:16 for TikTok, Reels, Shorts or ads and export.
Veo 3 release date.
Google DeepMind released Veo 3 in May 2025 at Google I/O, alongside a faster, cheaper Veo 3 Fast variant. It followed Veo 2 (late 2024), and its headline feature — native audio generated with the picture — made it the most talked-about model release of the year.
Its successor, Veo 3.1, launched in October 2025 with reference images, first-and-last-frame control, 4K output and scene extension. On Veo Studio it's now the default model for every plan, and Veo 3 remains available as an option.
Veo 3 specifications.
- Clip length
- 4–8 seconds
- Resolution
- 720p · 1080p
- Aspect ratios
- 16:9 · 9:16
- Inputs
- text · image (first frame)
- Audio
- native: music, SFX, dialogue
- Provenance
- SynthID invisible watermark
Veo 3.1 is here.
Veo 3.1 adds up to 3 reference images, first-and-last-frame control, 4K output and scene extension up to ~148 seconds, on every plan, at no extra cost. It's live now.
Veo 3 FAQ.
What is Veo 3?+
Veo 3 is an AI video generation model from Google DeepMind. It turns a text prompt or an image into a finished video clip with camera direction and native sound — music, effects and lip-synced dialogue are generated in the same pass as the picture.
When was Veo 3 released?+
Veo 3 was released by Google DeepMind in May 2025 at Google I/O, together with the faster Veo 3 Fast variant. Its successor, Veo 3.1, launched in October 2025 and is now the default model on Veo Studio.
Is Veo 3 free?+
You can use Veo 3 for free on Veo Studio's free plan, which includes a couple of 720p videos a month with no card required. Paid plans from $14/month add more videos, 1080p and 4K output, and a commercial licence.
How do I access Veo 3?+
Sign up for a free account and pick Veo 3 from the model selector in the prompt box (Veo 3.1 is the default). It runs in the browser, so it works on any device without an app or download.
Is Veo 3 open source?+
No. Veo 3 is a proprietary model from Google DeepMind, available through hosted apps and APIs rather than as open-source weights. Every Veo clip carries SynthID, Google's invisible provenance watermark.
Is Veo 3 available in the USA?+
Yes. Veo Studio runs Veo 3 in the browser and is available in the United States and worldwide, you just need an account.
Start creating with Veo 3.
The free plan gets you a couple of videos a month, enough to see what Veo 3 does with your idea.
Make your first video