Key Notes
- Google has extended Omni video generation to fifteen seconds and reduced its listed price for 720p output by 30%, making longer clips cheaper to produce.
- Conversational editing and reference guidance aim to preserve characters, products and scenes across revisions, with improvements measured against the previous version.
- The model is generally available through the Gemini API and Google Cloud, while access and allowances inside individual Google apps follow separate rollouts.
Google has released Gemini Omni 1.2 Flash, an update to its AI video model that makes clips longer and lowers the cost of producing them. Announced on November 5, the release supports videos up to 15 seconds and cuts the listed price of 720p generation by 30% to $0.07 per second.
The update also puts more emphasis on refining a clip through successive instructions while keeping characters, objects and scenes recognizable. Google’s announcement positions the release for repeat production work, including product videos, explainers and advertising variations.
Longer Clips at a Lower Starting Price
The Gemini API bills video output by token usage, and its pricing page estimates about $1.04 for a 15-second 720p clip. The advertised $0.07 per second is rounded. Inputs are charged separately, and additional attempts, edits or upscaling can increase the total project cost. For teams producing many variations, the number of usable results can matter as much as an individual render’s price.
The Cloud documentation specifies generation and editing durations from three to 15 seconds, with landscape and portrait aspect ratios of 16:9 and 9:16. There is a narrower limit for generation guided by both a first and last frame: those clips run from three to ten seconds. The maximum therefore depends on the workflow selected.
Google recommends drafting at 360p or 720p and upscaling the strongest results to 1080p or 4K. That gives creators a way to test the idea before spending more effort on the final version. A longer clip also provides more room for a short sequence to develop without stitching several separately generated shots together.
Editing the Same Scene Through Conversation
The model card describes text, images, audio and video as inputs, with video and synchronized sound as the output. Creators can use references to guide the result rather than relying on a written description alone. That is particularly relevant when a recognizable character, garment or product needs to survive a change in camera angle.
Conversational editing lets a creator request another background, lighting setup or visual treatment while continuing from an existing clip. The practical appeal is continuity: an advertiser changing the setting should not have to accept an unrelated performance or a different-looking product as the price of that revision. Whether the model preserves those details on a particular brief still needs checking.
The release builds on the workflow introduced with the original Omni launch. The progression is toward a repeatable production tool, where references and revisions remain useful across several steps. A compelling first result matters less if every small change forces the creator to rebuild it.
Better Consistency in Google’s Comparisons
Google DeepMind publishes comparative results against Omni 1.1. In its text-to-video visual-consistency comparison, the new version received a 62% win rate versus 38% for the older model. A reference-to-video identity-consistency comparison favored the update 63% to 37%.
Those are company-published, side-by-side results, not an independent ranking of every competing video model or a guarantee of a clean result on every prompt. Google also says fast sports action, rapid script changes and complex physics remain areas under development. The comparisons support a narrower claim: the company reports improvements over its own previous release.
API Availability and App Rollouts Differ
Gemini Omni 1.2 Flash is generally available on the paid Gemini API tier and through Google Cloud, while availability inside individual apps follows separate rollouts. Google says 15-second creation is reaching Google AI Ultra subscribers in the Gemini app, alongside uses in Flow, Google Vids and Google Ads. API pricing should therefore be distinguished from the features and allowances included in a particular app subscription.
Developers using Omni 1.1 also have a migration date to consider. Google says API requests to that version will automatically route to Omni 1.2 Flash from December 3, 2026, and be billed at the newer model’s rates. Teams with existing video workflows have a reason to test the update before that change.
The Cloud specification also lists support for generated speech, music and sound effects, as well as C2PA Content Credentials. Google says outputs carry its SynthID watermark. As AI video tools become easier to integrate into regular workflows, the useful measure is how reliably they deliver an editable, publishable result within the available budget.
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.
