Google DeepMind Launches Nano Banana 2 AI Image Model With 4K Support and Character Consistency

Google DeepMind launched the original Nano Banana in August 2025, followed by Nano Banana Pro in November. With strengths in character consistency, text rendering and image editing, the models targeted the generative AI market and competed with OpenAI's GPT Image series for developer and enterprise demand.

Google launched Nano Banana 2, also known as Gemini 3.1 Flash Image, on February 26, 2026. It supports resolutions of up to 4K and can maintain consistency across five characters and 14 objects in a single scene. The model topped the Artificial Analysis and LM Arena text-to-image rankings and placed third for editing. At about $0.07 per image, it costs roughly half as much as Nano Banana Pro and GPT Image 1.5.

Google is positioning Nano Banana 2 Lite as a streamlined image-generation model for enterprise automation workflows, targeting high-volume production of product images, advertising assets and internal content. Computing costs and response latency often constrain enterprise adoption of generative AI, making speed, affordability and scalability key factors in whether businesses deploy such models.

Google’s newly unveiled Nano Banana 2 Lite can generate an image in about 4 seconds and costs just $0.034 per 1,000 images, half the price of the previous offering. It is aimed at high-frequency, large-batch enterprise applications. Google also previewed Gemini Omni Flash, a multimodal video-generation model. Reports did not provide exact release dates or formal launch schedules for either product.

With users’ explicit consent, Google’s Personal Intelligence can connect to Gmail, Google Photos, YouTube and Google Search, allowing Gemini to generate or edit images based on their photos, interests and personal context. The integration makes prompting easier, while placing account-data permissions and privacy controls at the heart of the service’s expansion.

Google announced on June 29, 2026, that eligible US users with personal accounts could immediately access personalized image-generation features powered by Nano Banana in the Gemini app for free. Once Google Photos is connected, the system can select suitable images directly without requiring users to upload them each time. Paid Google AI subscribers can also use the more advanced Nano Banana Pro model.

Google launched the first Nano Banana model in August 2025, followed in November by Nano Banana Pro, which emphasized professional-grade quality. The new Nano Banana 2, powered by Gemini 3.1 Flash Image, aims to combine Pro-level image quality and reasoning with Flash speed, lowering the barrier to large-scale image creation and iterative editing for businesses.

Google released Nano Banana 2 on February 26, 2026, pricing image output at $60 per million tokens, half Nano Banana Pro’s $120. A 4K image costs about $0.151. The model supports resolutions from 512px to 4K, text generation within images and multilingual translation. It is being rolled out across Gemini, Search, AI Studio, Vertex AI, Flow and Google Ads.

Google DeepMind introduced Nano Banana in August 2025 and followed with Nano Banana Pro, which emphasized professional controls, in November. The new generation brings Pro’s world knowledge, reasoning and high-resolution capabilities to the faster Flash architecture, lowering the barrier for everyday users to generate and edit high-quality images.

Google released Nano Banana 2 on Feb. 26, 2026, under the formal model name Gemini 3.1 Flash Image. It improves detail, subject consistency and multilingual text generation, with support for resolutions from 512px to 4K. The model is available to free Gemini users and has been integrated into Google Search, Chrome and Flow, where it costs 0 credits to use.