Immediately, we’re releasing Veo 3.1 and Veo 3.1 Quick in paid preview within the Gemini API. This up to date mannequin presents a number of enhancements, in addition to improved outputs when producing video from pictures. These new fashions can be found by way of the Gemini API in Google AI Studio and Vertex AI. Veo 3.1 can be obtainable within the Gemini app and Movement.
Veo 3.1: our improved video era mannequin
Veo 3.1 and Veo 3.1 Quick empower builders to create extra partaking content material by means of important upgrades. The fashions now generate richer native audio, from pure conversations to synchronized sound results, and supply higher narrative management with an improved understanding of cinematic types. Enhanced image-to-video capabilities guarantee higher immediate adherence whereas delivering superior audio and visible high quality and sustaining character consistency throughout a number of scenes.
Along with the mannequin replace, we’re introducing new mannequin capabilities, together with utilizing reference pictures to information video era, prolong present Veo movies, and generate transitions between a primary and final body.
Information era with ‘Elements to video’
Now you can information the era course of by offering as much as 3 reference pictures of a personality, object, or scene. That is useful for sustaining character consistency throughout a number of photographs or making use of a particular model to your video.
from google import genai
from google.genai import varieties
shopper = genai.Shopper()
operation = shopper.fashions.generate_videos(
mannequin="veo-3.1-generate-preview",
immediate=immediate,
config=varieties.GenerateVideosConfig(
reference_images=[reference_image1, reference_image2],
),
)Python
See extra within the documentation.
Create longer movies with ‘Scene extension’
Your story is not restricted by the unique era. With Scene extension, you may create longer movies, even lasting for a minute or extra, by producing new clips that hook up with your earlier video. Every new video is generated primarily based on the ultimate second of the earlier clip. This maintains visible continuity, making the characteristic supreme for extending photographs with background audio.
from google import genai
shopper = genai.Shopper()
operation = shopper.fashions.generate_videos(
mannequin="veo-3.1-generate-preview",
immediate=immediate,
video=video_to_extend
)Python
See extra within the documentation.
Management your video with ‘First and final body’
Create clean, pure scenes that bridge two completely different pictures. By offering a beginning and an ending picture, you may direct Veo 3.1 to generate the transition between them, full with accompanying audio.
from google import genai
from google.genai import varieties
shopper = genai.Shopper()
operation = shopper.fashions.generate_videos(
mannequin="veo-3.1-generate-preview",
immediate=immediate,
picture=first_frame,
config=varieties.GenerateVideosConfig(
last_frame=last_frame,
),
)Python
See extra within the documentation.
What individuals are constructing with Veo 3.1
Promise Studios, a GenAI film studio, makes use of Veo 3.1 inside its MUSE Platform to boost generative storyboarding and previsualization for director-driven storytelling at manufacturing high quality.
Latitude is experimenting with Veo 3.1 in its generative narrative engine to immediately convey user-created tales to life.
Begin constructing immediately
Veo 3.1 and these new capabilities are actually obtainable in paid preview by way of the Gemini API.
- Dive into the documentation for detailed parameter data and video size management,
- Begin utilizing Veo 3.1 and the brand new options in Veo Studio, our new AI Studio demo app (paid Gemini API key required)
- Or leap proper into the code with the up to date cookbook information
Veo 3.1 is similar worth as Veo 3. We’re extremely excited to see what you’ll construct with these new capabilities!
