Google Releases Gemini 2.5 Flash Image Integrated with Adobe and OpenRouter

Google has officially launched the Gemini 2.5 Flash Image model, codenamed “nano-banana,” aimed at enhancing image generation capabilities. The new model improves image quality and control while emphasizing character consistency, multi-image fusion, and semantic editing capabilities. Notably, it marks the first collaboration with Adobe and OpenRouter, broadening its application ecosystem.

The development of Gemini 2.5 Flash began with small-scale testing at the end of last year, leading to an expansion of experiments globally through Google AI Studio in March of this year. The current release represents its official version.

Gemini 2.5 Flash Image has four main features: it ensures consistency in the appearance of the same character or product across different scenes; allows users to modify images precisely using natural language commands, such as blurring backgrounds, removing blemishes, or even changing poses; enables multi-image fusion for developers to combine several images, effortlessly placing objects into new scenes; and integrates world knowledge to interpret hand-drawn diagrams and educational applications, addressing complex editing needs.

The new model is accessible via the Gemini API and Google AI Studio at a cost of $30 for every million output tokens. Each image requires approximately 1,290 tokens, averaging around $0.039 per image. Additionally, all generated and edited images will come with a SynthID digital watermark for transparency and traceability.

Adobe has announced the integration of the Gemini 2.5 Flash Image model into Adobe Firefly and Adobe Express. Firefly is Adobe’s generative AI platform for image and design launched in 2023, while Adobe Express is a widely accessible all-in-one content creation platform. Adobe indicates that creators on social platforms, marketers, and visual designers can leverage this model to realize their creative ideas.

Adobe users can generate up to 20 images for free, while paid subscriptions unlock unlimited usage.

Furthermore, Gemini 2.5 Flash Image has become the first image-generating model on OpenRouter, claiming to assist in generating website pages, creating character-consistent comic storyboards, and devising numerous variations of logo designs.

OpenRouter, founded in 2023, serves as a model intermediary platform, acting as an App Store for AI models. It allows developers to access various AI models through a single API interface and currently supports over 480 models across text, code, and image generation.