Google Developer News: Project Genie, Android Studio Otter 3, Translate Gemma & Google I/O 2026

DeepMind Project Genie

• 🚀 Project Genie is an experimental research prototype from DeepMind that allows users to create and explore interactive 3D worlds.
• 💡 Powered by Genie 3, Nanobanana Pro, and Gemini, it generates environments in real-time as users interact.
• ✍️ Core capabilities include world sketching (text/image prompts), world exploration (navigating in first/third person), and world remixing (building on existing prompts).
• 🌍 Access is available for Google AI Ultra subscribers in the US (18+).

Android Studio Otter 3 Enhancements

• 🤖 Android Studio Otter 3 now allows users to choose any LLM for AI assistance, including OpenAI, Anthropic, local models via Olama, and Gemini.
• ✍️ Journeys enables natural language descriptions for end-to-end UI tests, with AI performing actions and summarizing results.
• 🎨 Deeper integration of Gemini in the compos preview streamlines UI development, from code generation to debugging.
• 🖼️ Features include creating UI from design mocks and matching UI to target images with suggested code.

Gemma Ecosystem and Genkit Go Updates

• 🌐 Translate Gemma introduces a new collection of open translation models built on Gemma 3, supporting 55 languages across various parameter sizes (4B, 12B, 27B).
• 🗣️ Trained on 55 core language pairs and nearly 500 additional pairs for high-quality performance.
• 🛠️ Genkit Go version 1.4.0 is available, featuring experimental sessions for multi-turn conversations and a refactored tools interrupts API.
• 📊 Prompt evaluations are now live in the Genkit dev UI for bulk testing and side-by-side comparisons.

Chrome 145 and Google I/O 2026

• 💻 Chrome 145 is now stable, introducing new CSS styling capabilities like column wrapping and extended support for customizable select lists.
• 🗓️ Google I/O 2026 is announced for May 19th and 20th at the Shoreline Amphitheater in Mountain View, California.