Google has unveiled Gemini 2.5 Flash Image, a cutting-edge image generation and editing model designed to simplify complex tasks while maintaining high-quality results. This update builds upon the success of native image generation in Gemini 2.0 Flash, addressing user requests for improved image quality and creative control.
The Gemini 2.5 Flash Image model is now available via the Gemini API and Google AI Studio for developers, with pricing starting at $30 per 1 million output tokens. This price point is consistent across all modalities on input and output.
One of the key features of this model is its ability to maintain character consistency across multiple prompts and edits. Developers can easily create apps using the Google AI Studio “build mode” that enable users to upload images, apply filters, and edit characters with precise control. The model also excels at adhering to visual templates, making it suitable for tasks such as real estate listing cards, uniform badges, or dynamic product mockups.
Gemini 2.5 Flash Image can perform targeted transformations using natural language, allowing users to blur backgrounds, remove stains, and alter subjects with ease. Additionally, the model benefits from Gemini’s world knowledge, unlocking new use cases in areas like image understanding and semantic editing.
The model’s capabilities are demonstrated through interactive template apps built in Google AI Studio. These tools enable users to merge multiple input images, create photorealistic fused images, and generate high-quality images with a single prompt.
OpenRouter.ai has partnered with Google to bring Gemini 2.5 Flash Image to its 3M+ developers, while fal.ai will make the model available to their broader developer community.
All images generated or edited with Gemini 2.5 Flash Image will include an invisible SynthID digital watermark, allowing users to identify and verify the AI-generated content.
Google is actively working on improving long-form text rendering, character consistency, and factual representation in the model. Users can provide feedback through the developer forum or X to help shape the future development of Gemini 2.5 Flash Image.
Source: https://developers.googleblog.com/en/introducing-gemini-2-5-flash-image