Product signal

Google launches Gemini Omni 1.1 Flash

Google has rolled Gemini Omni 1.1 Flash out to eligible Flow subscribers globally, adding longer scene extension, video references, and faster low-resolution previews to make video editing more iterative.

By All We Need Editorial System

Google has rolled out Gemini Omni 1.1 Flash to eligible Google AI Plus, Pro, and Ultra subscribers using Flow globally. Compared with Gemini Omni Flash’s May introduction to Flow, the update turns video editing more clearly into a workflow built around extension, reference inputs, and repeated previews at different resolutions.

Editing capability and availability both expand

Google says Gemini Omni 1.1 Flash can extend scenes to 40 seconds and use video references of up to three seconds as editing inputs, with output available at 1080p or 4K. This is more than a limited demonstration: Google says the release is available to eligible Flow subscribers globally. Gemini Omni Flash entered Flow in May as a precise editing tool across more than 140 countries. The new change advances the available inputs, duration, and output tiers together.

Cheaper previews change the iteration mechanism

The practical cost of video generation depends not only on a final render but on how many attempts are needed to correct framing, movement, and composition. Google says 360p previews can be up to 60% faster and cost one-third as much as 720p previews. Combined with reference inputs and scene extension, that enables lower-cost testing before higher-resolution output. Google DeepMind’s human-preference and benchmark-based evaluations provide methodological context for the company’s quality and instruction-following claims.

The competitive focus may shift to editing loops

The consequence is to place model capability inside a generate-review-adjust-render loop rather than merely improve a single prompt’s output. Faster previews may raise the amount of iteration users can afford and make resolution tiers part of product design. The strongest countercase comes from Google’s own model card: the model can still struggle with full consistency across multiple edits, complex motion, and accurate text rendering, so broader functionality does not yet establish reliable substitution for professional video production.

What to watch next

The next observable evidence is whether Google releases Flow usage, preview-to-final-render conversion, or creator-retention data; whether Omni 1.1 Flash reaches APIs or enterprise creative tools; and whether independent tests reproduce gains in multi-turn consistency, complex motion, and text rendering. Improvement on those measures would strengthen the case that editing loops, rather than one-shot generation, are becoming the key video-model battleground.

Sources