Google's latest update to Gemini Omni 1.1 Flash transforms AI video generation from a one-shot tool into a directed creative process, giving developers precise control over scene composition, character consistency, and video length. The update, announced by Google DeepMind product managers Anish Nangia and Alisa Fortin, introduces four major features designed to move the model from experimental demos into real production pipelines.

What Are the Four New Features in Gemini Omni 1.1 Flash?

The original Gemini Omni Flash model had a hard 10-second ceiling on every clip, which made it nearly impossible to tell coherent stories or build longer sequences. The 1.1 release addresses this and three other critical gaps that stopped teams from using the model in actual production work.

• Scene Extension: You can now continue generating footage from where a previous clip ended, in 10-second increments, up to a cumulative total of 40 seconds. The model maintains visual consistency and narrative flow across extensions, so shots can carry on or branch into new directions without visible seams.
• Keyframe Control: Specify the starting and ending frames of a shot, and Omni 1.1 generates the continuous video between those two points. This is ideal for complex camera orbits, zoom transitions, or seamless looping clips, turning "generate a video and hope the camera does something sensible" into a directed creative process.
• Character and Style Reference: Reference up to three seconds of video when crafting a new scene so the model holds onto visual context and keeps characters consistent across shots. Google's demo showed characters swapped into dance reference clips while maintaining one continuous shot with no scene cuts.
• Resolution Drafting: Generate lightweight previews in 360p resolution, which Google describes as up to 60% faster and a third of the cost of standard 720p, then upscale the final output to 1080p or 4K. This lets teams iterate cheaply on rough cuts before spending real compute on the final version.

How Much Does Gemini Omni 1.1 Flash Cost to Use?

There is no free tier for Gemini Omni 1.1 Flash. Every second of video you generate costs money from the first clip. The model bills video output at a fixed rate of 5,792 tokens per second of 720p video, which translates to approximately $0.10 per second at the current API pricing of $17.50 per 1 million tokens.

In practical terms, a full 10-second 720p clip costs around $1.00, while a 5-second clip runs about $0.50. The real cost advantage comes from the 360p draft tier, which costs roughly $0.03 per second, or about a third of the 720p rate. This means you can prototype a 40-second sequence for just a few cents, then pay the full rate only on the final version you upscale to 4K.

Google notes there is no Batch, Flex, or Priority tier for this model, unlike some other Gemini offerings, so there is no documented batch discount available. The pricing structure rewards teams who iterate on drafts before committing to final renders.

Where Is Gemini Omni 1.1 Flash Available?

Google shipped Omni 1.1 across multiple surfaces on launch day, signaling its intent to position the model as production infrastructure rather than a one-off demo tool. The model is available through Google AI Studio for direct API testing, the Gemini Enterprise Agent Platform for enterprises building on the Agent Platform API, Google Flow for all Google AI Plus, Pro, and Ultra subscribers globally, and the Gemini app where scene extension is live for the same subscriber tiers.

The technical specifications include a 131,072-token input context and 57,920-token output capacity, a 10-second maximum per generation request (extended through scene extension), support for up to three videos per prompt, 16:9 and 9:16 aspect ratios, and resolutions from 360p through 4K. Generated video carries C2PA Content Credentials for provenance tracking.

What Do Industry Partners Say About the Update?

"With extensions, richer reference material, and 4K resolution, Gemini Omni Flash takes teams beyond generating videos to truly directing them," said Itay Schiff, creative director at Figma Weave.

Itay Schiff, Creative Director at Figma Weave

Google's launch partners, who are already running the model in production, highlighted how the new controls shift the creative paradigm. The update moves from a generative tool that produces unpredictable results to a directed tool where creators define the boundaries and the model fills in the motion and details within those constraints.

However, independent developers on Hacker News raised a practical concern: the 360p draft previews are only useful if the final 720p render actually matches them. Since video generation is non-deterministic, meaning the same prompt can produce different results each time, a draft may not accurately predict what the final output will look like. This gap between preview and final render could limit the efficiency gains the draft tier promises.

How Does Omni 1.1 Flash Compare to Other AI Video Models?

Google's AI lead Logan Kilpatrick positioned Omni Flash as state-of-the-art in video editing at $0.10 per second, matching the pricing of Veo 3.1 Fast. However, independent AI analyst Rohan Paul suggested the real product shape emerges when chaining multiple models together rather than relying on any single model alone. He noted that Nano Banana 2 Lite can generate reference images, which Gemini Omni Flash then animates, creating a more powerful workflow than either model in isolation.

The 1.1 update reflects Google's strategic commitment to video generation as a core capability. While OpenAI abandoned Sora entirely, Google is doubling down on production-ready controls and developer tooling, positioning Omni as infrastructure for creative teams rather than a consumer novelty.