Monitor this situation.

Unsubscribe anytime.

[SITUATION] · [QUIET] · [TECHNOLOGY]

2 clusters · 25 sources · 5 days · First seen 25 days ago · Last updated 20 days ago

Google Nano Banana image generation model development

Overview

The Google Nano Banana model family serves as a tool for prompt-led image creation, allowing users to generate and edit visuals through text-based instructions. These models facilitate rapid prototyping by enabling tasks such as background changes, lighting adjustments, and scene restyling.

Google DeepMind has released Nano Banana 2.1, an updated version based on Gemini 3.6 Flash designed to serve as an efficient counterpart to the Nano Banana Pro model. This iteration introduces significant improvements in visual design, natural appearance, and precise text rendering, alongside better accuracy for infographics. Technical enhancements include the ability to process up to 14 reference images to maintain the consistency of up to four characters and ten objects. The model also features a ‘Thinking’ feature with minimal, medium, and high levels to influence image quality.

Nano Banana 2.1 supports various output resolutions, including 1K, 2K, and 4K, and is being rolled out across the Gemini app, Google Search’s AI Mode, Google Ads, Google AI Studio, and developer tools such as Flow and Stitch. New capabilities include mask-based editing, which allows users to modify specific areas of an image without regenerating the entire scene, and fixes for tiling artifacts in panoramic formats.

While the new model reduces the cost of standard 1K image generation to approximately $0.0336—roughly half the cost of Nano Banana 2—input token costs for text, images, and video have increased to $1.50 per million tokens. Nano Banana 2 is expected to be phased out by October 29, 2026.

Entities

Google · Gemini 3.6 Flash · Alphabet · Google DeepMind · Nano Banana 2.1

Claims

What the coverage asserts, and how many sources carry each claim.

[● 18 SOURCES] Google launched the Nano Banana 2.1 image generation and editing model on October 6, 2026.
[● 16 SOURCES] Nano Banana 2.1 supports image generation and editing at 1K, 2K, and 4K resolutions.
[● 11 SOURCES] The API cost for a standard 1K image is $0.0336, which is approximately half the price of the previous Nano Banana 2 model.
[● 9 SOURCES] The model can process up to 14 reference images and maintain consistency for up to four characters and ten objects.
[● 7 SOURCES] Nano Banana 2.1 is based on the Gemini 3.6 Flash model.
[● 7 SOURCES] The model includes configurable Thinking levels: minimal, medium, and high.
[● 2 SOURCES] Nano Banana 2 is scheduled to be shut down on October 29, 2026.

Timeline

20 days ago

[TECHNOLOGY] 24 sources
Google launches Nano Banana 2.1 AI image model
Google launched Nano Banana 2.1, an AI image model featuring improved subject consistency, 4K support, and mask-based editing. API costs for image generation have been reduced by roughly half.
25 days ago

[TECHNOLOGY] 2 sources
Google Nano Banana models drive rise in prompt-led image creation
Prompt-based image generation tools, such as Google’s Nano Banana model family, are streamlining creative workflows through text-to-image and image-to-image editing capabilities.

Sources

360mozambique.com · ad-hoc-news.de · ascii.jp · ceotech.it · criptonoticias.com.br · decrypt.co · dev.to · gamemag.it · kulturegeek.fr · meeco.kr · mobilebit.com.br · mobilissimo.ro · news.yesky.com · prtimes.jp · punto-informatico.it · seroundtable.com · smartworld.it · stadt-bremerhaven.de · the-decoder.de · theinspiringjournal.com · thetechnologyexpress.com · tokenpost.com · unwire.hk · weel.co.jp · world-today-news.com

This summary has been updated 2 times: see revision history