Google has officially released Nano Banana 2 (technically named Gemini 3.1 Flash Image), the latest iteration of its viral AI image generation model.
The goal is straightforward: combine the high-fidelity output of Nano Banana Pro with the speed of the Gemini Flash architecture, at a significantly lower cost.
The most anticipated upgrade is Subject Consistency, which allows the model to maintain the appearance of up to five characters and track up to 14 objects within a single workflow — making it far more practical for storyboarding, comics, and brand content creation.
Advanced World Knowledge lets the model draw from Gemini’s knowledge base and real-time web search to generate more accurate infographics and data visualizations.
Precision Text Rendering has also improved notably, now supporting multi-language text generation and translation within images, which is useful for localized advertising.
Output resolution ranges from 512px up to 4K with multiple aspect ratio options.
On pricing, Nano Banana 2’s Flash-tier image output costs roughly half that of Pro mode, opening up near-Pro quality generation to a wider audience. It has also replaced Nano Banana Pro as the default image engine across all Gemini app modes, though Google AI Pro and Ultra subscribers can still switch back to the Pro model via the menu.
For content authenticity, all images generated by Nano Banana 2 carry an invisible SynthID watermark and are compatible with the C2PA Content Credentials standard. Google noted that SynthID verification has been used over 20 million times since launch.
Stay tuned to MyMobileTrend for more tech coverage
