blog.google
|
ksl
|
|
Google released Nano Banana 2, built on Gemini 3.1 Flash, and immediately made it the default image generator across the Gemini app, Google Lens, Flow, and AI Mode in 141 countries. The model handles up to 4K resolution, maintains consistency across five characters and fourteen objects per workflow, and renders text more accurately than its predecessor – all at Flash-tier speed rather than Pro latency. The original Nano Banana went unexpectedly viral, and Google is clearly capitalizing on that momentum by shipping the successor as infrastructure rather than an optional upgrade. With OpenAI, Midjourney, and Stability all iterating on their own generation pipelines, image models are converging on the same feature checklist: speed, resolution, consistency, and embedded provenance via watermarking.
