Overview
- Built on Gemini 3 Pro, the model is live in the Gemini app globally with a limited free quota before reverting to the original Nano Banana, while Plus, Pro and Ultra subscribers get higher limits.
- It adds precise photographic controls, localized edits and higher resolutions, generating 2K and 4K images, blending up to 14 inputs and keeping consistent likeness for up to five people.
- Google emphasizes accurate, legible in‑image text across multiple languages and can ground visuals in up‑to‑date facts by referencing Google Search for recipes, weather, sports and data visuals.
- Provenance measures include SynthID invisible watermarking for all outputs, a visible mark for free and many Pro‑tier generations, optional clean exports for AI Ultra and some developer tools, and an in‑app checker to verify if images were made with Google AI.
- Availability extends to Search AI Mode in the U.S. for AI Pro and Ultra subscribers, NotebookLM for subscribers, Flow for Ultra, Workspace apps like Slides and Vids, Google Ads, and developer access via the Gemini API, AI Studio and Vertex AI; TechCrunch reports per‑image pricing of about $0.139 for 1080p/2K and $0.24 for 4K, higher and slower than the original model.