Gemini Omni 1.2 Flash: longer video, lower pricing and the limits to check

Gemini Omni 1.2 Flash Explained

Gemini Omni 1.2 Flash is Google’s AI video model for generating clips and revising them through conversational instructions. The November 5, 2026 release focuses on longer shots, visual reference adherence and follow-up edits.

Teams can evaluate it through Google’s Cloud offering or Google Vids. The supported features, account limits and access requirements depend on the route they use.

Public record: Google launch announcement — November 5, 2026

Gemini Omni 1.2 Flash: What’s New

The release brings video generation and editing from 3 to 15 seconds, conversational revisions and improved adherence to visual references. Google Cloud also lists extension, upscaling, speech, music and sound effects.

Vids access follows its own rollout

The Workspace team says Omni 1.2 Flash is arriving directly in Google Vids, with a duration slider from 3 to 15 seconds. Its full rollout began November 5, with 1–3 days for feature visibility across Rapid Release and Scheduled Release domains.

The availability list includes personal Google accounts and multiple Workspace editions. There is no admin control for the feature; AI Expanded Access add-on users receive higher usage limits. Check the Vids limits for the account you intend to use before planning production volume.

Measure the accepted clip, then the next edit

Our suggested trial is a short product shot using assets you have permission to use. Define three things that must remain fixed, such as the silhouette, label and camera position. Generate the shot, then request one background change.

Review the beginning, middle and end, and compare the revision with the original. Record each attempt, elapsed time and charge. A cheaper generation is useful only if the accepted result fits the brief; a revision that changes an approved detail adds work elsewhere.

Keep the first trial narrow. Once the product survives one edit, add camera motion or dialogue as a separate test. That makes failures easier to diagnose and gives the team a practical basis for deciding where this release fits.

Public record: Google launch announcement — November 5, 2026 · Google Cloud model specifications — checked November 6, 2026 · Google Workspace Vids rollout — November 5, 2026

Gemini Omni 1.2 Flash Pricing and Availability

Google launched Gemini Omni 1.2 Flash on November 5, 2026. Its announcement describes conversational edits and improved adherence to visual references, alongside a 720p price of $0.07 per second, 30% below Omni 1.1.

At that announced rate, one 15-second 720p generation works out to $1.05. This is arithmetic for a single generation, not a quote for a finished campaign or a price for every output resolution.

Google Cloud lists gemini-omni-1.2-flash as generally available, with a November 5 release date. Its specification supports video generation, editing, extension and upscaling, plus speech, music and sound effects.

The $0.07-per-second figure applies to the announced 720p rate. Check the price and usage limits for your chosen resolution and access route before estimating a production budget. Google Vids follows its own account eligibility and usage limits.

Public record: Google launch announcement — November 5, 2026 · Google Cloud model specifications — checked November 6, 2026 · Google Workspace Vids rollout — November 5, 2026

Gemini Omni 1.2 Flash: Limitations and First Impressions

The documented generation and editing range is 3–15 seconds, but first-and-last-frame generation is limited to 3–10 seconds. Supported aspect ratios are 16:9 and 9:16. The page lists 360p, 720p, 1080p and 4K outputs.

Integration matters too: this Cloud model page lists the Interactions API as supported, while GenerateContent, Chat Completions and batch inference are not. An existing video integration may therefore need more than a model-name replacement.

Google also identifies fast sports action, rapid script changes and complex physics as areas still being improved. Its consistency claims should be evaluated on your own material; Promptwire has not independently tested the model.

Our initial assessment is that longer clips and a lower announced 720p rate make this release worth evaluating for short production workflows. The practical question is whether each revision preserves approved details. These are first impressions from the published specifications, rather than results from hands-on testing.

Public record: Google launch announcement — November 5, 2026 · Google Cloud model specifications — checked November 6, 2026