gemini omni 1.2 flash: 15-Second Clips and a 30% Price Cut

gemini omni 1.2 flash is Google’s generally available video model bump that stretches native clips to 15 seconds, adds multi-turn conversational edits, and cuts 720p pricing to $0.07 per second. Google’s November 5 developer blog says Omni 1.2 Flash is GA on the Gemini API and Gemini for business, lifts native video from 10s to 15s, enables multi-turn editing, and is 30% cheaper than Omni 1.1 at that $0.07/sec 720p rate. The Verge’s same-day piece says the model generates 3–15 second clips, edits without full re-generation, and improves continuous camera moves and dialogue.

What actually happened

gemini omni 1.2 flash is a creative-AI ship in the same 24-hour window as OpenAI’s cyber model and Anthropic’s finance tool — which is why it belongs in memcool’s AI lane even if you never touch a prediction market. Primary documentation is Google’s own blog; The Verge supplies secondary product framing on clip length and editing behavior.

Fifteen seconds is still short versus Hollywood, but it is a meaningful jump from a 10-second ceiling for social and ad workflows that need a continuous take. Multi-turn conversational edits — changing a clip with follow-up prompts without full re-gen — is the feature that matters for iterative creators who previously paid to regenerate entire takes.

Pair this launch with Google’s education-AI footprint in google schools ai and the broader lab-tooling race in cathedral systems openai. Different surfaces, same week: labs shipping concrete SKUs, not just keynote vapor.

Why a 30% price cut travels as far as the extra five seconds

Creative APIs are sold on unit economics as much as on demos. Dropping to $0.07/sec at 720p with an explicit 30% cut versus Omni 1.1 is Google telling developers the model is meant to be used in loops, not only in once-a-day showcases. Cheaper iterative edits plus multi-turn prompting is how you get app builders to wire video into products.

The Verge’s notes on continuous camera moves and dialogue are quality claims from launch coverage. Treat them as reviewer/reporter framing of Google’s improvements, not as your own side-by-side lab test unless you ran one.

Also keep the GA surfaces straight: Gemini API and Gemini for business per Google’s blog. If a thread says the model is only in an unlisted research preview, it is behind the official GA language.

How to read 3–15 second clips without overselling

A 15-second native clip is a building block, not a feature film. Long-form pipelines will still stitch, extend, or script around short takes. The win is fewer hard cuts and less full-regen waste when a client says "make the camera drift left."

Safety and likeness issues do not vanish because the clip got five seconds longer. Google’s blog is a capability and pricing announcement; it is not a license to ignore consent and deepfake policy. Stay inside the product terms.

Competitive timing without invented benchmarks

We are not reprinting rival VBench scores we did not fetch for this batch. The fair same-day frame is qualitative: Google shipped longer native clips and cheaper 720p while other labs shipped cyber and finance SKUs. Readers can care about all three without forcing a single winner chart.

What to do as a reader

If you build on Gemini video, read the official Google Omni 1.2 Flash blog for GA surfaces, clip limits, and the $0.07/sec 720p number before you rewrite your billing code. Use The Verge as a secondary explainer, not as the pricing source of record.

gemini omni 1.2 flash is a creative-model update, not a token. Nothing here is investment advice.

Timestamp discipline applies across this whole desk: every percentage, volume print, and product claim above is tied to the dated sources in our research pack. If a screenshot in your chat disagrees with the primary link, the primary link wins. Refresh live pages before you act on a stale mid, and do not invent fills, unlocks, or revenue figures that are not in those sources. Nothing in this article is investment advice, financial advice, or a promise of profit — it is a same-day briefing for readers who want confirmed context before the crowd finishes arguing about screenshots.