# Google Omni 1.2 Flash > Google Omni 1.2 Flash is a video generation and editing model that creates clips up to 15 seconds and supports multi-turn editing to refine the same video. It is for advertisers and creators making product videos or short clips who need to iterate... - Type: AI tool - Page: https://completeaitraining.com/ai-tools/google-omni-12-flash/ - Published: 2026-11-06 - Updated: 2026-11-08 - Author: Jeroen Erne - Jobs: Creatives, Marketing - Topics: Generative-Video, Video-Editing, Text-To-Video - Reading time: 2 min ### About Google Omni 1.2 Flash Google Omni 1.2 Flash is Google DeepMind's video generation and editing model, now generally available. It produces clips up to 15 seconds and supports a draft-then-upscale workflow, letting users iterate at lower resolutions before finalizing at 1080p or 4K. The model is accessible through the Gemini API, Google Cloud, Flow, and Vids. ### Review Omni 1.2 Flash builds on its predecessor with longer clip lengths, lower per-second cost, and a multi-turn editing capability that doesn't require regenerating the entire video. The model targets short-form production workflows where refining a single clip over several rounds matters more than generating a one-off output. What follows is a breakdown of what the tool actually does, where it fits, and where it doesn't. #### Key Features - **Multi-turn editing without regeneration.** Users can swap a product, change lighting or weather, switch the aspect ratio, or modify backgrounds across multiple editing rounds on the same clip. The model preserves characters, products, and styles across those turns, though drift may accumulate with repeated edits. - **15-second clip generation with adjustable length.** Clips can be set from 3 to 15 seconds, up from the 10-second ceiling in Omni 1.1. The model supports continuous camera moves like FPV flythroughs, zoom-outs, and orbits. - **Draft-then-upscale workflow.** Users draft at 360p or 720p, make edits, then upscale to 1080p or 4K. This keeps iteration costs low while preserving a path to higher-resolution output. - **Input flexibility.** The model accepts character images, product images, style references, short motion clips, audio cues, first and last keyframes, and voice references. - **SynthID watermarking.** All outputs carry an imperceptible SynthID watermark for provenance tracking. #### Pricing and Value 720p generation costs $0.07 per second, a 30% reduction from Omni 1.1 pricing. Google Vids users with a Google account or Workspace subscription can generate 15-second clips at no extra cost. Google AI Ultra subscribers get access through the Gemini app. For API usage through Google AI Studio and Vertex AI, the model ID is gemini-omni-1.2-flash. Pricing for 1080p and 4K upscaled outputs has not been separately detailed. #### Pros - Multi-turn editing keeps the core clip intact while allowing iterative changes to products, backgrounds, and lighting. - Lower per-second cost and a draft-at-low-resolution workflow make multiple takes and revisions economically practical. - Wide input support-keyframes, audio cues, voice references, product images-gives editors several ways to guide output. - Access points span API, cloud, and consumer-facing tools like Google Vids and Flow, covering both programmatic and manual workflows. - SynthID watermarking is built into the output pipeline rather than being an opt-in step. #### Cons - Character and product identity may drift across multiple edit turns, similar to repeated image edits, and Google hasn't published quantitative benchmarks on fidelity after four or five rounds. - Upscaling to 4K is available, but the quality gap between the 720p draft and the final upscaled output hasn't been independently characterized. - Not well suited for projects requiring clips longer than 15 seconds or for teams that need frame-level editing precision rather than prompt-and-refine iteration. Omni 1.2 Flash fits teams producing ads, product videos, or short social clips who revise a single video over several rounds instead of starting from scratch each time. The draft-then-upscale approach and multi-turn editing address a specific pain point-losing a good take to one wrong detail-but the model's 15-second ceiling and potential for accumulated drift mean it won't replace dedicated compositing tools for longer or more precise work. --- This is the Markdown form of https://completeaitraining.com/ai-tools/google-omni-12-flash/. The whole site is mapped at https://completeaitraining.com/llms.txt.