Google makes Gemini Omni 1.2 Flash generally available with 15-second clips

Google’s new video model cuts costs by 30% and extends native clip length to 15 seconds, aiming to speed professional video workflows. The claims are self-reported and lack independent verification.

By Mira Sethi

November 5, 2026 at 2:03 PM EST2 min read

Key takeaways (3)

1.Audio references need allowlisting; Gemini app access is rolling out to AI Ultra subscribers; Gemini for business is "coming soon."
2.Google's blog lists Gemini for business as generally available; Google Cloud's post says it is coming soon.
3.Neither Google post includes independent test results; invideo's "evals" remark is a customer statement without published data.

Key takeaways (3)

1.Audio references need allowlisting; Gemini app access is rolling out to AI Ultra subscribers; Gemini for business is "coming soon."
2.Google's blog lists Gemini for business as generally available; Google Cloud's post says it is coming soon.
3.Neither Google post includes independent test results; invideo's "evals" remark is a customer statement without published data.

Google said Nov. 5, 2026 that its Gemini Omni 1.2 Flash video model, which generates clips up to 15 seconds, is generally available on the Gemini API and Google Cloud.

Google's two announcements differ on one surface. The blog post 1 says the model is generally available across the Gemini API and Gemini for business. The Google Cloud post 2 says business users can create with it in Gemini for business and Google Vids, but its getting-started list reads "Try it in Google Vids; coming soon to Gemini for business."

Google said the model is priced at $0.07 per second 1 for 720p video, a 30% drop from Gemini Omni 1.1, with similar savings at other resolutions according to Google Cloud.

Background

The Gemini Omni 1.2 Flash release is compared throughout with Gemini Omni 1.1. Google said clips can now run up to 15 seconds, up from 10 seconds with Omni 1.1 1. Duration can be set in one-second increments from 3 to 15 seconds.

The blog also says that "for the first time, Gemini Omni is now generally available." Neither Google post says whether Omni 1.1 Flash remains on offer.

What Google says changed

Google lists several separate improvements. It says the model has better reference adherence 1, keeping faces, garments, packaging typography and brand colors consistent across camera angles. It also describes multi-turn conversational editing, "more fluid, natural-sounding dialogue, clearer lip-sync, and distinct multi-speaker conditioning," and camera moves with "significantly fewer unwanted jump cuts."

The blog also describes stronger first-to-last frame consistency. Its demonstrations include a 4-shot, 15-second macro explainer with accurately spelled callouts and a Pantheon flight with "zero jump cuts."

Google said users can draft at 360p or 720p, make follow-up edits instead of regenerating a clip, and upscale to 1080p or 4K. Provisioned Throughput is available, and outputs carry an imperceptible SynthID watermark.

Access limits and customer comments

Google Cloud's post says audio references are gated behind allowlisting 2. The blog says Google Vids clip generation comes at no additional cost with a Google account or Google Workspace subscription. It also says 15-second video in the Gemini app is rolling out to Google AI Ultra subscribers 1.

The Google Cloud post carries quotes from Artlist, FLORA, invideo, Krea, Picsart, Photoroom and QuickFrame AI. Artlist's Roee Peled said the model helps customers move "from experimentation to professional content production at scale." Photoroom's Matt Rouif said 15-second video was "a top user request."

Invideo's Shivam Rastogi said the company ran the model "through our evals" and called it "one of the best we've tested." The post does not publish those evaluations. The post presents Runway through a description of its products and a showcase of what it made with the model, without a quote.

Neither Google post includes independent test results. The blog's comparisons between 1.1 and 1.2 are Google-chosen examples with no stated methodology.

Google said development is ongoing on fast sports action, rapid script changes and complex physics.

Sources2 documents

1. blog.google

blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/

1 checked claim

2. cloud.google.com

cloud.google.com/blog/products/ai-machine-learning/gemini-omni-1-2-flash-generally-available

Citations link to the documents this article draws on; they do not imply independent corroboration.

NVIDIA adds four open-weight Cosmos 3 Nano checkpoints for interactive simulation

The models target robotics and autonomous-driving researchers who need controllable, photorealistic environments to train and evaluate policies without real-world risk.

Mira Sethi · 20 minutes ago

NVIDIA says full-stack codesign raises AI inference output per HBM byte

NVIDIA's Thursday post frames AI memory as an efficiency problem across the AI factory's hierarchy and claims up to 30× higher tokens per dollar on Vera Rubin NVL72 versus GB300 NVL72, attributing gains to full-stack codesign rather than raw HBM capacity increases.

Mira Sethi · about 4 hours ago

Mistral releases Large 4 in preview and expands Studio platform

Mistral AI launched Mistral Large 4 (Le Chonk) and expanded its Studio enterprise platform to help businesses build, deploy and govern AI agents, workflows and applications across multiple environments.

Mira Sethi · about 5 hours ago