Google DeepMind releases TxGemma, built on Gemma, which predicts therapeutic properties, and Agentic-Tx, powered by Gemini 2.0 Pro, which tackles complex research problem-solving with advanced tools.

The Google I/O 2025 puzzle used the Gemini API to generate dynamic riddles for bonus worlds, enhancing player engagement and scalability. Here's what our developers learned on using the Gemini API effectively, including creativity, design, and implementation strategies.

Gemma 3 is a new, advanced version of the Gemma open-model family featuring multimodality, longer context windows, and improved language capabilities, with various sizes and deployment options for developers to experiment.

The experimental native image generation feature of Gemini 2.0 Flash – allowing for the combination of text and images, conversational image editing, and leveraging real-world knowledge for contextual visuals – is now available for developers to test through Google AI Studio and the Gemini API.

Gemma 3 1B, a new small language model for mobile and web applications via Google AI Edge, is now available, with increased efficiency, improved performance, and offline availability.

ShieldGemma 2, built on Gemma 3, is a 4 billion parameter model that can be used as an input filter for vision language models or an output filter for image generation systems, and is designed to respond to a wide range of diverse and nuanced imagery.

An enhancement to Google Cloud Dataflow templates for MongoDB Atlas enables direct integration of JSON data into BigQuery, eliminating complex data transformations, reducing operational costs, and enhancing query performance for users.

Google Cloud Next 2025, happening April 9-11 in Las Vegas, will feature expanded developer content, interactive experiences, and opportunities to connect with peers and Google experts.

A new experimental Gemini Embedding text model, now available in the Gemini API, achieves top rankings on the Massive Text Embedding Benchmark (MTEB) leaderboard and offers expanded language support and high-dimensional embeddings.

This blog post introduces Gemini's code execution feature, which allows the AI model to generate and run Python code for tasks like solving equations, data analysis, and creating visualizations.

CalCam, a calorie-tracking app, uses the Gemini API to analyze meal photos, providing users with fast and accurate nutritional information. Polyverse, CalCam's creator, highlights Gemini API's speed, accuracy, and structured JSON output are crucial for CalCam's seamless user experience and efficient development, allowing for easy integration and detailed food analysis.

The Data Science Agent in Google Colab, powered by Gemini, can now generate complete, working notebooks from simple natural language descriptions, so developers can automate data analysis tasks, saving time to focus on deriving insights.

Gemini 2.0 Flash-Lite is now generally available in the Gemini API for production use in Google AI Studio and for enterprise customers on Vertex AI. 2.0 Flash-Lite offers improved performance over 1.5 Flash across reasoning, multimodal, math and factuality benchmarks. For projects that require long context windows, 2.0 Flash-Lite is an even more cost-effective solution, with simplified pricing for prompts more than 128K tokens.

PaliGemma 2 mix, an upgraded vision-language model, is now available, offering capabilities like image captioning, OCR, and object detection in various sizes.

Langbase empowers developers to build and deploy powerful, scalable AI agents by leveraging the Google Gemini API, particularly Gemini 1.5 Flash, unlocking a new era of intelligent applications and streamlined workflows.

A practical guide to constructing a Gemma 2-based Agentic AI system – a type of AI that can make its own decisions and use external tools to achieve goals – that can generate dynamic content for a fictional game world.

Google I/O returns May 20-21. Watch the livestreams for updates on Android, AI, web, and cloud. Registration is open on the Google I/O website.

Imagen 3 – now available in Google AI Studio and the Gemini API – offers developers state-of-the-art image generation with brighter, better-composed images in diverse styles, and simplified image generation through text prompts.

The Gemini 2.0 model family is seeing significant updates, including the release of Gemini 2.0 Flash, which is now production-ready and boasts higher rate limits, enhanced performance, and simplified pricing. Developers can also start testing an updated experimental version of Gemini 2.0 Pro today. Additionally, a new variant called Gemini 2.0 Flash-Lite, specifically designed for large-scale workloads, will be made available next week.

The Gemini API and ESP32 microcontroller simplify custom voice commands for IoT devices, leveraging speech recognition for devices to understand and react to custom commands, bridging the gap between digital and physical worlds.