Build the next generation of visual apps with gemini-2.5-flash-image-preview in @googleaistudio. Explore these demo apps and remix them to make them your own 🧵↓
@googleaistudio The “Past Forward” demo app shows how the model can maintain a subject’s identity throughout different environments and edits. Fork the demo to build your own app with more control for a huge range of applications.
@googleaistudio The “PixShop” demo app allows you to make precise changes with simple text commands. Remix the app to use the model to remove elements from a photo, alter poses, and change element colors.
@googleaistudio The "Gemini Co-Drawing" app demonstrates the multimodal model's ability to read hand-drawn diagrams, perform calculations, and follow complex editing instructions. Remix the app to build creative solutions that can reason across both text and images to generate content.
@googleaistudio “Home Canvas” showcases the model’s creative composition strength. Go beyond single-image edits and create new scenes quickly by dropping in product images. Remix the app to build tools that blend elements from multiple images into a single scene.
@googleaistudio Remix and customize these demo apps or build your own using Gemini 2.5 Flash Image in Google AI Studio: aistudio.google.com/apps
• • •
Missing some Tweet in this thread? You can try to
force a refresh
We’re excited to share about MedGemma, SignGemma, and DolphinGemma 🧵↓
1️⃣MedGemma is a new collection of Gemma 3 variants created to accelerate the development of medical text and image applications, such as analyzing medical images or clinical reasoning.
MedGemma's 4B (multimodal) and 27B (text-reasoning) models offer strong performance for medical text/image comprehension, making them adaptable across medical applications. MedGemma runs efficiently on a single GPU for fine-tuning and inference, giving you full health data control thanks to its open nature. developers.google.com/health-ai-deve…