Google has broadly released its Gemini 2.5 Flash AI image editing model, providing developers with advanced capabilities for image generation, natural language photo editing, and complex image manipulation, accessible via the Gemini API and Google Cloud's Vertex AI. The model features a pay-as-you-go pricing structure for production use ($0.039 per image) and is being integrated into platforms like Adobe Firefly, underscoring Google's strategic focus on building flexible AI ecosystems. This move intensifies competition in the AI media generation sector, notably against releases such as OpenAI's Sora 2.
eWeek content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More Google has rolled out Gemini 2.5 Flash, its advanced AI image editing model, making the tool widely available. Its native image generator app, Nano Banana, now has 10 new aspect ratios and the option for image-only output. Aspect ratios enable AI-generated images to be created with the ideal shape for various uses, such as vertical social media posts or sweeping landscape shots. Gemini 2.5 Flash Image now supports landscape (21:9, 16:9, 4:3, 3:2), square (1:1), portrait (9:16, 3:4, 2:3), and âflexibleâ (5:4, 4:5) aspect ratios. The other new feature is the ability to specify image-only output. The model defaults to returning text and image responses, where a text explanation accompanies the image output, but this can be overwritten by replacing responsemodalities=[âTextâ, âImageâ] with responsemodalities=[âImageâ]. See Gemini 2.5 Flashâs features in pre-made apps, or create your own Gemini 2.5 Flash was initially released as a preview version via the Gemini API and Google AI Studio at the end of August. Developers can use it to create their own AI image editing apps or apply some of the demo ones that Google has vibe coded. The powerful model can create a character and place it in various scenes, such as a desert or underwater, while maintaining a consistent look. Google demonstrates this in its Past Forward app, which can take a picture of the user and then place them in different decades, and Fit Check, which can change a personâs clothes or pose. Gemini 2.5 Flash can interpret natural language commands for photo editing. In the Pixshop app, users can explicitly ask it to remove people, blur backgrounds, or colourise photos without any technical knowledge of editing techniques. Meanwhile, Gemini Co-Drawing allows them to draw and then ask AI to refine it through a text prompt. While it is primarily an image model, Gemini 2.5 Flash also has real-world knowledge that can be used in its apps. Gemini Co-Drawing can interpret what you are drawing and provide additional information, such as by adding labels, measuring angles, or inserting relevant objects. Users can also merge multiple images with the new model. In the Home Canvas app, for example, they can drag and drop an object image onto a scene, and the system will generate a photorealistic composite that blends the object naturally into its surroundings. Developers can try it out for free in Google AI Studio All these pre-existing apps are available to try out in Google AI Studio, and the Build section allows users to build their own apps based on Gemini 2.5 Flash with a simple text prompt. The Gemini API lets the model be integrated into existing code, allowing enterprise users to access it through Google Cloudâs Vertex AI platform. The Gemini 2.5 Flash Image is accessible for free through Google AI Studio; however, for production use or high-volume needs, developers must switch to pay-as-you-go pricing. Native image generation costs $0.039 per image, while text and multimodal output are priced at $30 per million tokens. Elsewhere, OpenRouter has added Gemini 2.5 Flash as its first-ever image model, fal.ai is bringing it to its developer community, and Adobe has integrated it into Firefly. This reflects a broader strategic shift among AI companies: away from competing on raw performance and toward offering flexible ecosystems where users can pick the best tool for the job. Gemini 2.5 Flash may not be the biggest AI media generator to be released this week. OpenAIâs Sora 2 has arrived, allowing users to create videos featuring copyrighted characters. Google has broadly launched Gemini 2.5 Flash, an advanced AI image editing model, now widely available with new native image generation features like 10 aspect ratios and image-only output. This model empowers developers to create sophisticated applications, leveraging capabilities such as consistent character generation across diverse scenes and natural language photo editing, allowing users to modify images without technical expertise. The strategic release highlights Google's commitment to enhancing its AI ecosystem, evidenced by its inclusion in themes like Artificial Intelligence and Technology & Innovation. Gemini 2.5 Flash is integrated into several platforms, including Adobe Firefly, and is accessible via Google AI Studio and Google Cloud's Vertex AI for enterprise users. While free for developers in AI Studio, production use is subject to a pay-as-you-go model, priced at $0.039 per image for native generation and $30 per million tokens for text/multimodal output. This pricing structure and broad integration underscore Google's strategic shift towards providing flexible AI ecosystems rather than solely competing on raw performance. The launch occurs within an increasingly competitive AI media generation market, notably alongside OpenAI's Sora 2 release. Google's strong positive sentiment (0.8 for GOOGL/GOOG) and the market's optimistic tone regarding this product launch suggest a favorable reception to its advanced features and ecosystem approach. However, Adobe's lower sentiment score of 0.4, despite integrating the model, indicates potential nuances in how the market views its position within this evolving landscape.
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Request a DemoOverall Sentiment
strongly positive
Sentiment Score
0.75
Ticker Sentiment
