What changed from 1.1, where the consistency and dialogue gains show, and how to prompt Google's newest video model.
Short answerGemini Omni Flash 1.2 is Google's newest video model, and it is now available in Recraft Studio. Compared to 1.1, it keeps characters and products steadier from the first frame to the last, follows camera directions with fewer unwanted cuts, and produces cleaner lip-synced dialogue. In Recraft Studio, clips run up to 10 seconds with native audio.
Google released Gemini Omni Flash 1.2 on November 5, 2026, as a direct update to 1.1. Google's own name for it is Gemini Omni 1.2 Flash. Most of the changes are about control rather than new features.
Characters, objects, and scenes drift less over the course of a clip, camera moves follow the prompt with fewer unwanted cuts, and reference images for a character, product, or style carry through more faithfully. On Google's own API the model can also run longer clips. In Recraft Studio, generations are up to 10 seconds.
| Gemini Omni Flash 1.1 | Gemini Omni Flash 1.2 | |
|---|---|---|
| Consistency across a clip | More drift and morphing | Less drift, steadier faces and products |
| Camera moves | More unwanted cuts | Follow the prompt with fewer jump cuts |
| Dialogue | Native audio with speech | Clearer lip-sync, distinct voices for several speakers |
| References | Character, product, and style references | Closer adherence to the references |
| Clip length in Recraft Studio | Up to 10 seconds | Up to 10 seconds |
| Aspect ratios | 16:9 and 9:16 | 16:9 and 9:16 |
Model changes come from Google's announcement on November 5, 2026. Clip length is the Recraft Studio limit.
The improvements matter most in clips where something has to stay the same while everything else moves. A product can rotate through a full turn and keep its label and proportions. A character can walk toward the camera and still have the same face when they arrive.
Dialogue is the other clear step. Two characters can trade a line each with mouths that match the words, which is usually the first thing to break in a short generated scene.
Google is direct about where the model still struggles: fast sports action, rapid script changes, and complex physics. If your brief depends on those, plan for more takes.
Google's prompting guide for Gemini Omni Flash comes down to being explicit about the things the model would otherwise decide for you.
.svg)
[0-3s].With 10 seconds to fill, write the clip as two or three beats rather than one long description. That gives each beat enough time to land.
Gemini Omni Flash 1.2 sits in the model picker next to the other video models in Recraft Studio, on plans that include video and external models. Clips run up to 10 seconds. It joins Gemini Omni Flash, which arrived in Recraft Studio in July.
It works in the same workspace as Recraft V4.1 and your images, so the key visuals and the motion for a campaign stay in one project.
In Recraft Studio, up to 10 seconds per clip. Google's own API allows longer clips for this model, but Recraft Studio generations are capped at 10 seconds.
Gemini Omni Flash 1.2 keeps characters, products, and scenes steadier across a clip, follows camera directions with fewer unwanted cuts, and produces clearer lip-synced dialogue, including scenes with several speakers. Reference images for a character, product, or style also carry through more faithfully.
Yes. Every clip is generated with native audio, including speech with lip-sync and scenes with several speakers. Describe the sound you want in the prompt, or write "No dialogue" if you don't want speech.
No. Gemini Omni Flash 1.2 is available in Recraft Studio on plans that include video and external models, and each generation uses credits. Current plans are listed at recraft.ai/pricing.
