Google DeepMind Ships Nano Banana 2, Its Most Capable Image and Video Generation Model Yet

Nano Banana 2, built on Gemini Flash, combines state-of-the-art image generation, editing, storyboarding, and real-time web integration in a single model — and it's rolling out widely.

Google DeepMind announced the launch of Nano Banana 2, its latest multimodal generation model built on the Gemini Flash architecture. As @GoogleDeepMind detailed in a thread, the model handles image creation, editing, storyboarding, and video generation within a single system — a notable departure from the piecemeal approach most competitors still use, where different models handle different modalities.

Unlock the full briefing

Get every story in today's briefing, the full archive, and the daily AI intelligence brief.

All stories today

Full archive

Daily brief

Cancel anytime. Payments powered by Stripe.