- Google released Gemini 2.5 Flash Image, its latest AI tool for image generation and editing.
- The tool allows image editing and character consistency, aiming to rival OpenAI‘s offerings.
- Developers can customize and deploy image applications instantly with built-in templates in Google AI Studio.
- Gemini 2.5 Flash Image is available through OpenRouter and fal.ai, expanding access to global users.
- All AI-generated images are marked with an invisible SynthID watermark and metadata for transparency.
Google launched Gemini 2.5 Flash Image on Tuesday, expanding its AI product suite with a tool that can create and edit images with a high degree of accuracy and consistency. The release, available through Google apps and platforms, is aimed at narrowing the gap with OpenAI, which has led in the AI image generation field.
Gemini 2.5 Flash Image allows users to modify visuals using natural language, such as changing poses or combining multiple images, while preserving details like faces and backgrounds. According to a Google blog post, the system can place the same character in different environments and show products from multiple angles without altering the main subject.
The model, which previously appeared under the name “nano-banana” on the testing site LMArena, supports complex editing tasks. “State-of-the-art image generation and editing” and “incredible character consistency” are available, Google AI Studio said on social media. Integration with AI Studio enables developers to remix templates and launch custom image projects quickly.
Google states that this model can interpret diagrams, blend reference materials, and leverage outside knowledge within a single prompt. The cost for using Gemini 2.5 Flash Image is $30 per million output tokens, which is about four cents per image, via Google Cloud. The model is also offered through OpenRouter and fal.ai for broader distribution.
The company reported 400 million monthly active Gemini users in August 2025, while OpenAI surpassed 700 million weekly active users after releasing GPT-4o with image support in March 2025.
Google includes an invisible SynthID watermark and metadata with all outputs to indicate that images are AI-generated, addressing ongoing concerns about authenticity and possible misuse.
✅ Follow BITNEWSBOT on Telegram, Facebook, LinkedIn, X.com, and Google News for instant updates.
