Google's Gemini Omni Flash is now available in Pencil. Generate video from text or images, and for the first time, edit existing videos with just a prompt and a reference image.
✨What's new
Text-to-Video: Generate video from a prompt, just like our other video models
Image-to-Video: Use an image as a start frame, an end frame, or a reference (one of these at a time, they can't be combined)
Video editing and guidance: Give the model a video plus a prompt, and optionally a reference image. Example: upload a clip and a mascot image and prompt "replace the animal in the video with the mascot in the image"
⚙️How to turn it on
If your workspace already has any Google video model enabled, Gemini Omni Flash is already on
Otherwise an admin can enable it under AI Governance, then Settings, then Video Generation
👥Who it's for: Anyone producing or iterating on video who wants to tweak footage with a prompt instead of regenerating from scratch
🟢Live: today, (In preview from Google)
📋Good to know: Video references up to 10 seconds (retry if a reference is ignored), 70 MB limit on video uploads, one video reference at a time.
🚫Not doable: Voice editing or uploaded audio references.
