In the present day, we’re launching three massive Veo updates: assist for vertical format outputs (9:16 facet ratio), 1080p HD output, and new, decrease pricing. We’re additionally making Veo 3 and Veo 3 Quick steady and prepared for scaled manufacturing use within the Gemini API.
Extra Veo, for much less
Efficient at the moment, we’re making it extra reasonably priced to generate high-quality movies with Veo 3 and Veo 3 Quick with worth reductions throughout the board.
- Veo 3: Now $0.40 / second; was $0.75 / second
- Veo 3 Quick: Now $0.15 / second; was $0.40 / second
Vertical video and 1080p assist
With Veo 3 and Veo 3 Quick, now you can set the facet ratio to 9:16 so that you could generate vertical format movies which might be excellent for mobile-first and social use circumstances. Simply set the ‘aspectRatio’ parameter to ‘9:16’ and also you’re able to go.
You can even generate Veo 3 and Veo 3 Quick movies in 1080p HD, permitting you to create even larger high quality by setting the ‘decision’ parameter to ‘1080p’.
To indicate how one can construct new experiences, try the MediaSim demo app in Google AI Studio. It integrates the tldraw SDK for an interactive canvas demonstrating how Gemini’s multimodal understanding, era and picture enhancing 🍌 capabilities might be mixed with Veo 3 video and audio era capabilities to create multimodal media simulations.
Get began constructing with Veo 3 in minutes
With this launch, we’re making it extra reasonably priced to construct functions that assist prime quality video generations. We’re excited to see what builders will create with Veo 3 and Veo 3 Quick!
Right here’s a fundamental Python instance to create a video (ensure that to make use of the most recent google-genai model):
import time
from google import genai
from google.genai import varieties
shopper = genai.Consumer()
operation = shopper.fashions.generate_videos(
mannequin="veo-3.0-fast-generate-001",
immediate="a close-up shot of a golden retriever taking part in in a area of sunflowers",
config=varieties.GenerateVideosConfig(
negative_prompt="barking, woofing",
aspect_ratio="9:16",
decision="720p",
),
)
# Ready for the video(s) to be generated
whereas not operation.accomplished:
time.sleep(20)
operation = shopper.operations.get(operation)
print(operation)
generated_video = operation.response.generated_videos[0]
shopper.recordsdata.obtain(file=generated_video.video)
generated_video.video.save("golden_retriever.mp4")
Python
Prompts
1: The shot is of a lone rock climber skillfully making her ascent up a snow-capped mountain at dawn. Capturing her as she navigates a piece. The digital camera is shut only a few toes away. She seems to be on the digital camera with a smile {of professional} admiration. Wiping her forehead with a gloved hand, she seems to be immediately into its lens and says with an impressed, enthusiastic tone, “Veo 3 is now like 50% cheaper and better high quality so go construct” earlier than turning to proceed her climb.
2: Generate a excessive decision video, shot in ultra-slow movement with a macro lens. The video captures a single drop of magical ink falling into crystal clear water. As an alternative of merely clouding the water, the ink blossoms right into a miniature, residing, ethereal diorama of nature, full of animals and flowers, earlier than gently fading out
3: A cinematic time-lapse video of a single seed sprouting, rising right into a lush plant, after which blooming with vibrant flowers, all set towards a subtly altering background that hints on the passage of seasons. Give attention to the intricate particulars of the plant’s improvement and the interaction of sunshine and shadow







