Gemini Omni 1.2 Flash is generally available with cheaper video output (November 5, 2026)

What you get

gemini-omni-1.2-flash is a stable video model in the Gemini API. It makes videos of 3 to 15 seconds from text and images, with native audio, at 360p up to 4K. You can edit a generated video by talking to it, and extend it up to 30 seconds in total. It keeps characters and subjects more consistent across up to 6 reference images.

Video output costs $12.00 per 1M tokens, down from $17.50 on Omni 1.1 Flash. Input is $1.50 per 1M tokens and text output $9.00. Paid-tier rate limits are doubled. There is no free tier.

From December 3, 2026, requests to gemini-omni-1.1-flash go to gemini-omni-1.2-flash automatically and are billed at 1.2 Flash rates.

Prerequisites

• A paid-tier Gemini API key in GEMINI_API_KEY. The REST example below reads it from API_KEY, so set that too or swap the name.
• The Google Gen AI Python SDK for the Python example.

Steps

Python, from the official guide:

import base64
from google import genai

client = genai.Client()

interaction = client.interactions.create(
    model="gemini-omni-1.2-flash",
    input="A marble rolling fast on a chain reaction style track, continuous smooth shot."
)
with open("marble.mp4", "wb") as f:
    f.write(base64.b64decode(interaction.output_video.data))