Nano Banana 2.5 Is Here: Google's Fast AI Image Model, Upgraded

Google's Nano Banana line has been the default answer for fast, high-quality AI image generation ever since the original dropped in August 2025. Nano Banana 2 (Gemini's image model) kept the crown for speed, edit consistency and price. Now Nano Banana 2.5 is here — the mid-cycle refresh Google has been quietly testing, and it's already climbing the image leaderboards.

What is Nano Banana 2.5?

Nano Banana 2.5 is Google's latest image generation and editing model, sitting between the raw power of its flagship models and the velocity of the Flash line. It was spotted in testing under the codename "spicy-mayo" on the image arenas, where it quickly rivalled the best closed models for photorealism and instruction-following edits.

What's new

• Better text rendering — signs, labels and poster-style typography finally survive generation without looking mangled.

• Stronger instruction edits — multi-part edit prompts ("remove the bag, keep the lighting") land far more reliably than on Nano Banana 2.

• Photoreal skin and materials — fewer plastic-looking faces, more natural fabric and reflections.

• Same conversational workflow — edits stack in a chat, so you refine an image the way you'd brief a designer.

Where it beats the competition

Against GPT Image 2.5, Nano Banana 2.5 trades blows on quality but keeps Google's traditional advantage: speed and cost per image. Against Midjourney and Flux it wins on editability — you can talk to it, iterate on one region, and keep everything else stable. If you make product shots, UGC-style creatives or social assets in volume, that combination is hard to ignore.

Try Nano Banana 2.5 on Automatica

Automatica gives you Nano Banana 2.5 in a clean, fast interface — generate, edit and iterate without writing a single prompt like a spell. No subscriptions eating your credits overnight, just pay for what you create.

Create with Nano Banana 2.5 on Automatica soon →