Gemini Omni 1.1 Flash is Google's model for generating and editing video from text, images or existing clips, steered through plain conversation. Scene extension is the headline change, stretching a clip to 40 seconds while keeping characters and lighting consistent. Through the API, a second of 720p video runs about $0.10. Developers reach it in Google AI Studio, subscribers in the Gemini app and Flow.
Scene extension in Gemini Omni 1.1 Flash analyzes up to ten seconds of the source footage before generating what comes next, against a single second for Veo. Extensions chain in ten-second steps up to 40 seconds of continuous story, without the main character swapping jackets halfway through.
The rest of the control desk follows suit. You can pin the first and last frames of a shot to sketch a camera move, or drop in up to three seconds of real footage as a style reference. Conversational editing stays on board too, you type a request and the shot regenerates, closer in spirit to AI video editing tools than to a bare prompt box.
The 360p draft mode renders up to 60% faster and costs a third of a standard 720p pass. The workflow is deliberate, test ten cheap variations, then upscale only the keeper to 1080p or 4K.
That rate also matches Veo 3.1 Fast, billed around $0.10 per second at 720p. Rates move with each Google announcement, so the Gemini API pricing page stays the reference before any production run.
| Resolution | Role in the workflow | Price per second |
|---|---|---|
| 360p | Drafts and quick tests | $0.03 |
| 720p | Standard generation | $0.10 |
| 1080p | Web delivery | $0.15 |
| 4K | Upscaled final render | $0.30 |
Gemini Omni 1.1 Flash is available in Google AI Studio through the Gemini API, in Flow and on the Gemini Enterprise Agent Platform, while Google AI Plus, Pro and Ultra subscribers get scene extension inside the Gemini app. Adobe has plugged the model into Firefly as well.
Every generated clip carries the invisible SynthID watermark, checkable from the Gemini app, Chrome or Google Search (handy the day a suspicious video makes the rounds at the office). Voice and speech editing remains switched off for now, a stated guardrail against deepfakes.
Not through the API, which bills from the first second, around $0.10 at 720p and $0.03 for 360p drafts. Google AI Plus, Pro and Ultra plans include the model in the Gemini app and Flow through monthly credits, and the first Omni Flash also rolled out at no cost on YouTube Shorts.
Omni 1.1 Flash leads on conversational edits, mixed references and long scene extension, while Veo 3.1 is often favored for first-pass cinematic polish. Both fast tiers sit around $0.10 per second, so running the same brief through each and comparing accepted takes remains the fair test.
Yes, every clip embeds SynthID, Google DeepMind's invisible marker, detectable through the Gemini app, Gemini in Chrome and Google Search. Press testing has also shown the model can produce characters resembling famous franchises, which grants no right to publish or monetize such clips, so check likeness and trademark questions before commercial use.
Google is retiring the preview version and points developers to Gemini Omni 1.1 Flash, which keeps the same request format and the same per-second price. Existing integrations mostly swap one model identifier and carry on.
Verdict: Forty seconds of consistent footage and three-cent drafts make the pitch on their own. Ad studios, marketing teams and developers building video features get an engine made for fast iteration, with 4K upscaling saved for the final cut.
