IMPLICATOR
Google DeepMind released Gemini Omni 1.1 Flash on Aug. 27, raising the limit for generated video sequences from 10 seconds in the previous release to 40 seconds. The model now reads up to 10 seconds of prior footage before adding a fixed increment, rather than only the final one second. That gives developers more control over continuity while keeping each new generation short.
What Changed
- Gemini Omni 1.1 Flash raises generated video sequences from 10 seconds in the previous release to a cumulative 40 seconds, built in 10-second increments.
- The model now reads up to 10 seconds of prior footage before continuing a shot, rather than only the final second.
- A new 360p draft mode runs up to 60% faster than the standard 720p setting and is priced at one third, with finished output exporting at 1080p or 4K.
- Pricing is unchanged at roughly $0.10 per second of 720p video on the paid tier, and Google publishes no per-second rate for 4K.
AI-generated summary, reviewed by an editor. More on our AI guidelines.
Longer scenes
The original Gemini Omni Flash release arrived on May 19. Omni 1.1 adds 10-second segments that can be repeated until a sequence reaches the new cumulative limit, although a single generation still stops at 10 seconds.
Developers can also specify the first and last frames of a shot. The model generates the footage between those keyframes, a control intended for camera orbits, zoom transitions and loops. Video input can include up to three seconds of reference video to help preserve a character or visual style across shots.
Faster drafts
A new 360p draft mode is up to 60% faster than the standard 720p setting and is priced at one third of the standard setting. The speed claim is based on Google's own system throughput as of Aug. 27. Finished output can be exported at 1080p or 4K.
The Gemini API pricing page lists Omni 1.1 on the paid tier only. As of Aug. 27, 720p video is priced at roughly $0.10 per second, the same rate charged when the developer API launched on June 30. Billing uses 5,792 output tokens for each second of 720p video. Input is $1.50 per million tokens and video output is $17.50 per million.
The pricing page lists no per-second rate for 360p, 1080p or 4K, leaving the price of a 4K export undisclosed. Omni runs through Google's stateful Interactions API, which carries a previous video and its references into the next request.
Distribution and limits
The update is available through the Gemini API and Google AI Studio, as well as Google's enterprise agent platform. Google Flow offers it to AI Plus, Pro and Ultra subscribers worldwide, while scene extension is also reaching those subscribers in the Gemini app. WPP has integrated Omni Flash into WPP Open, and Adobe plans to add it to Firefly.
FREE WEEKDAY MORNING BRIEFING
Don’t miss the next AI story that matters.
The Implicator Morning Briefing filters the AI news cycle to the stories worth your attention and explains their consequences. From San Francisco, every weekday at 4:45 a.m. Pacific, 7:45 a.m. Eastern.
About five minutes. No hype. No spam.
Character consistency across edits and accurate text rendering remain open problems in the model card. Speech and audio editing also remain withheld beyond the avatar feature while the company tests how to release those controls responsibly. Every generated clip carries a SynthID watermark.
Know someone who'd find this useful? ✉️ Email it to a friend in one click, or they can subscribe free here.
Rival tools
ByteDance's Seedance 2.5, announced on June 23, can make clips up to 30 seconds long, export at 4K and accept up to 50 simultaneous reference inputs. Omni 1.1 still limits each generation to 10 seconds. OpenAI's Sora product is no longer available, and its Sora 2 video models and Videos API are deprecated and will shut down on September 24, 2026.
Hell Grind, a 95-minute AI-generated science fiction action film from Higgsfield AI, screened around Cannes in May 2026. It was made in roughly two weeks for about $500,000, of which about $400,000 went to AI compute, and it still required a 15-person team. The first 25 minutes alone took more than 16,000 initial generations, cut down to 253 final shots.
Ron Schmelzer, a Forbes contributor who has covered AI since 2018, wrote that Google's growing set of overlapping video products and entry points, including Veo, Veo 3.1 Lite, Flow, Google Vids, the Gemini app and now Omni, "could make the product story harder for users to follow."
Nicole Brichtova, a product management director at Google DeepMind, told TechCrunch in May that the original clip cap was a deployment choice, not a model constraint. A higher-end Omni Pro is planned without a release date and will arrive when Google sees "a step change above Flash."
Frequently Asked Questions
How long can a Gemini Omni 1.1 Flash video be?
A single generation still stops at 10 seconds. Those 10-second segments can be chained through scene extension until a sequence reaches a cumulative 40 seconds, up from the 10-second limit in the previous release.
What does Gemini Omni 1.1 Flash cost?
Roughly $0.10 per second of 720p video on the paid tier, billed at 5,792 output tokens per second. Input is $1.50 per million tokens and video output is $17.50 per million. That rate is unchanged from the June 30 developer API launch. Google publishes no per-second rate for 360p, 1080p or 4K.
What is the 360p draft mode for?
Prototyping. It runs up to 60% faster than the standard 720p setting and is priced at one third, so developers can iterate on a shot before rendering final output at 1080p or 4K. The speed figure is Google's own system throughput comparison.
Where can developers access the model?
Through the Gemini API and Google AI Studio, Google's enterprise agent platform, and Google Flow for AI Plus, Pro and Ultra subscribers worldwide. Scene extension also reaches those subscribers in the Gemini app.
How does it compare with rival video models?
ByteDance's Seedance 2.5 makes clips up to 30 seconds, exports at 4K and accepts up to 50 simultaneous reference inputs, while Omni 1.1 caps a single generation at 10 seconds. OpenAI's Sora 2 video models and Videos API are deprecated and shut down on September 24, 2026.
AI-generated summary, reviewed by an editor. More on our AI guidelines.
Marcus Schuler
Editor-in-Chief and founder of Implicator.ai. Former ARD correspondent and senior broadcast journalist with 10+ years covering tech. Writes daily briefings on policy and market developments. Based in San Francisco. E-mail: editor@implicator.ai
