Overview

• It is available today in the Gemini app for all users and to developers through the Gemini API, Google AI Studio and Vertex AI.
• The system supports multi‑turn editing, blends multiple photos into a new scene and applies styles across images while keeping people and pets looking consistent.
• Google says the model led evaluations on LMArena and other benchmarks after weeks of viral testing under the pseudonym “nano‑banana.”
• All images created or edited in Gemini carry a visible mark plus an invisible SynthID watermark, and Google’s terms prohibit generating non‑consensual intimate imagery.
• Adobe said it will offer Google’s model in Firefly and Express, underscoring growing competition with established image editors and rival AI models.