Today, we’re launching three big Veo updates: support for vertical format outputs (9:16 aspect ratio), 1080p HD output, and new, lower pricing. We're also making Veo 3 and Veo 3 Fast stable and ready for scaled production use in the Gemini API.
Link to Youtube Video (visible only when JS is disabled)
Effective today, we’re making it more affordable to generate high-quality videos with Veo 3 and Veo 3 Fast with price reductions across the board.
With Veo 3 and Veo 3 Fast, you can now set the aspect ratio to 9:16 so that you can generate vertical format videos that are perfect for mobile-first and social use cases. Just set the ‘aspectRatio’ parameter to ‘9:16’ and you’re ready to go.
You can also generate Veo 3 and Veo 3 Fast videos in 1080p HD, allowing you to create even higher quality by setting the ‘resolution’ parameter to ‘1080p’.
Link to Youtube Video (visible only when JS is disabled)
Link to Youtube Video (visible only when JS is disabled)
To show how you can build new experiences, check out the MediaSim demo app in Google AI Studio. It integrates the tldraw SDK for an interactive canvas demonstrating how Gemini’s multimodal understanding, generation and image editing 🍌 capabilities can be combined with Veo 3 video and audio generation capabilities to create multimodal media simulations.
Link to Youtube Video (visible only when JS is disabled)
Link to Youtube Video (visible only when JS is disabled)
With this release, we’re making it more affordable to build applications that support high quality video generations. We're excited to see what developers will create with Veo 3 and Veo 3 Fast!
Here’s a basic Python example to create a video (make sure to use the latest google-genai version):
import time
from google import genai
from google.genai import types
client = genai.Client()
operation = client.models.generate_videos(
model="veo-3.0-fast-generate-001",
prompt="a close-up shot of a golden retriever playing in a field of sunflowers",
config=types.GenerateVideosConfig(
negative_prompt="barking, woofing",
aspect_ratio="9:16",
resolution="720p",
),
)
# Waiting for the video(s) to be generated
while not operation.done:
time.sleep(20)
operation = client.operations.get(operation)
print(operation)
generated_video = operation.response.generated_videos[0]
client.files.download(file=generated_video.video)
generated_video.video.save("golden_retriever.mp4")All videos generated by Veo will continue to include a digital SynthID watermark. To get started, check out the documentation, a Veo 3 starter app in Google AI Studio, or our Veo cookbook.
Prompts
1: The shot is of a lone rock climber skillfully making her ascent up a snow-capped mountain at sunrise. Capturing her as she navigates a section. The camera is close just a few feet away. She looks at the camera with a grin of professional admiration. Wiping her brow with a gloved hand, she looks directly into its lens and says with an impressed, enthusiastic tone, "Veo 3 is now like 50% cheaper and higher quality so go build" before turning to continue her climb.
2: Generate a high resolution video, shot in ultra-slow motion with a macro lens. The video captures a single drop of magical ink falling into crystal clear water. Instead of merely clouding the water, the ink blossoms into a miniature, living, ethereal diorama of nature, filled with animals and flowers, before gently fading out
3: A cinematic time-lapse video of a single seed sprouting, growing into a lush plant, and then blooming with vibrant flowers, all set against a subtly changing background that hints at the passage of seasons. Focus on the intricate details of the plant's development and the interplay of light and shadow
Más allá de la propagación hacia atrás: el poder simbólico de JAX revela nuevas fronteras en computación científica
Introducing EmbeddingGemma: The Best-in-Class Open Model for On-Device Embeddings
Las extensiones A2A hacen posible la funcionalidad personalizada de agente
Google AI Edge Gallery: Ahora con audio y en Google Play