Google DeepMind’s Gemini app introduces the advanced “Gemini 2.5 Flash Image Generation” model, enhancing image editing capabilities significantly. This innovative model excels in prompt accuracy, outperforming previous versions like GPT-4o in generating precise edits while maintaining character consistency across various scenes. With features like localized adjustments, users can effortlessly blur backgrounds, erase objects, or modify colors using text prompts. The integrated PixShop template app showcases these editing functionalities in a user-friendly interface. Additionally, the model supports complex image compositions and style transfers, allowing seamless blending of multiple images and patterns. Gemini 2.5 Flash also incorporates real-world reasoning, visually illustrating cause-and-effect scenarios. Available for users via the Gemini app, the model can be accessed by selecting the “Flash” language option. The pricing structure for developers includes a $30 charge per million output tokens, ensuring efficient cost management for image processing and editing tasks.
Source link
Share
Read more