Gemini Image Editing: Every New Feature in 2026
Google has shipped image editing updates to Gemini at a fast pace throughout 2026 — new models, live camera editing, multi-image blending, and video generation. If you use Gemini regularly and have lost track of what's actually new versus what's been around for a while, here's a clear rundown of what's shipped, what each feature does, and what it means if you care about the watermark situation.
The headline updates
Part of a recent Gemini Drop rolled out across web, iOS, and Android, this lets you point your phone's camera and give spoken instructions while Gemini applies edits to the live feed in real time — powered by the Nano Banana image engine. Instead of taking a photo, editing it, then reviewing the result, the editing happens as you're framing the shot.
You can now upload two photos and have Gemini combine them into a single coherent image — putting yourself in a scene with a pet, merging a product photo with a background, or turning a photo into a custom figurine-style render. This builds on the original Nano Banana model's focus on maintaining a consistent likeness for people and pets across edits, rather than generating a new face each time.
Announced at Google I/O 2026, Gemini Omni blends text, image, audio, and video inputs to generate dynamic video content. In practice, this means you can take an image you've created or edited in Gemini and turn it into a short video — useful for product demos, social content, or just animating a static generation.
Rolling out first in the UK, Germany, France, Spain, and Italy, this lets you edit photos in Google Photos through natural conversation — saying something like "remove the reflections and fix the washed out colors" instead of using manual sliders. It's powered by Gemini models under the hood.
What hasn't changed: the watermark situation
Across all of these updates, one thing has stayed constant: Gemini-generated and Gemini-edited images still carry both the visible star watermark and the invisible SynthID signature. Google explicitly confirms this remains the policy regardless of which model or interface produced the image — Nano Banana, Nano Banana 2, or the new Live camera mode all apply the same disclosure approach.
This makes sense given the EU AI Act's Article 50 compliance deadline approaching in August 2026 — Google isn't going to roll back its watermarking infrastructure while adding new generation capabilities, it's the opposite incentive.
2026 Gemini image timeline at a glance
What this means if you're using Gemini images professionally
More editing power means more reasons to export a clean version of your work — for presentations, portfolios, or marketing material — without the visible star in the corner. None of these new features change the legal or practical situation around the visible watermark: if you generated or edited the image yourself, removing the visible mark for your own use remains straightforward and generally acceptable, as covered in our explainer on why Gemini adds the watermark in the first place.
The new Live camera and multi-image blending features produce the exact same watermark pattern as standard Gemini generations, so the removal process doesn't change — auto-detection for the star still works the same way regardless of which Gemini feature created the image.
Got a fresh Nano Banana 2 image to clean up?
Free Gemini watermark removal — auto-detection built in, no account needed.
Try WatermarkOff free →Related reading
Sources: Google Blog, Google Developers Blog, Google Cloud Blog, Nokia Power User (via Let's Data Science). This article tracks publicly announced Gemini features as of its publication date and will be updated as new releases ship.