AI & Creative Tools

Google's Gemini Omni 1.2 Flash Stretches AI Video to 15 Seconds and Cuts Pricing 30%

The update lands four months after the original Gemini Omni Flash launch, extending clip length and dropping the per-second price in the same release.

Google released Gemini Omni 1.2 Flash on November 5, a version update to the Gemini Omni Flash video model it introduced earlier this year. The headline changes are practical rather than flashy: longer clips and a lower price per second.

What actually changed

Gemini Omni 1.2 Flash extends the model’s maximum generated clip length to 15 seconds, up from the shorter ceiling of earlier Omni Flash releases, and cuts per-second generation pricing by 30%. Neither change is a new capability so much as a widening of what was already possible — longer sequences without stitching multiple generations together, at a lower running cost per finished clip.

For anyone building on top of the API rather than using it through a consumer surface, that combination compounds: a 15-second clip at 30% lower per-second pricing is a meaningfully cheaper total bill than the same duration would have cost under the prior pricing, independent of the length increase itself.

Versioning as the real story

This is the third numbered Omni Flash release inside Google’s generative-media lineup this year, following the original Gemini Omni Flash launch and the pairing with Nano Banana 2 Lite that Google shipped in July. The pattern that’s emerging is a model family on a fast, incremental release cadence — duration, latency, and price moving independently round to round — rather than one that waits for a single large capability jump to justify a version bump.

That’s worth naming because it changes how a creative-tools audience should read these releases: the interesting signal isn’t “is 1.2 better than 1.1,” it’s which specific dial (length, cost, fidelity) moved this time, and whether that’s the one that was blocking your actual workflow.

Where it sits in a crowded field

Gemini Omni 1.2 Flash ships into the same week as OpenAI’s Images 2.6 (Flare and Sunburst), which pushes on speed and SVG export rather than video duration — a reminder that the major labs are currently differentiating on different axes of the same problem rather than converging on one. For creative teams juggling multiple generative-video vendors, duration and per-second pricing are now genuinely comparable line items across providers, which they weren’t a year ago when clip lengths varied wildly and pricing models weren’t apples-to-apples.