Google

Gemini 3.5 Live Translate v3.5

Gemini 3.5 Live Translate is Google's latest audio model that delivers near real-time, natural-sounding speech-to-speech translation in over 70 languages. It automatically detects languages and generates smooth translations that preserve the speaker's intonation, pacing, and pitch, enabling fluid communication across language barriers.

LiteRT.js

LiteRT.js is a high-performance web AI runtime that enables developers to run machine learning models directly in the browser. By leveraging WebAssembly and hardware acceleration like WebGPU and WebNN, it allows for low-latency, private, and serverless AI inference using .tflite models.

DiffusionGemma v1.0

DiffusionGemma is an experimental open model developed by Google that utilizes text diffusion techniques to generate entire blocks of text simultaneously, achieving up to four times faster text generation on dedicated GPUs compared to traditional autoregressive models. Released under an Apache 2.0 license, this 26-billion parameter Mixture of Experts (MoE) model is designed for researchers and developers working on speed-critical, interactive local workflows such as in-line editing, rapid iteration, and generating non-linear text structures.

TranslateGemma v1.0

TranslateGemma is a suite of open machine translation models developed by Google AI, built upon the Gemma 3 architecture and designed to support 55 languages. The models are available in 4B, 12B, and 27B parameter sizes, optimized for deployment across various devices, from mobile and edge hardware to laptops and cloud instances.

Conductor v1.0.0

Conductor is an extension for Gemini CLI that facilitates context-driven development by enabling developers to create formal specifications and plans alongside their code in persistent Markdown files. This approach allows for planning before building, reviewing plans prior to coding, and maintaining developer control throughout the development process.

Gemini CLI v1.6 Preview vv1.6 Preview

Google updated Gemini CLI to v1.6 Preview with improved spatial reasoning capabilities, specifically targeting developers who use AI for hardware and robotics-related coding. The terminal-first agent mode now handles spatial concepts more accurately, making it a go-to tool for embedded systems and robotics software engineers.

Gemini 3.7 Flash v3.7 Flash

Gemini 3.7 Flash is an advanced, high-performance AI model optimized for coding, agentic workflows, and complex reasoning tasks. It delivers significant improvements in debugging, issue resolution, and web development while maintaining cost-efficiency for developers and enterprises.

Google AI Studio Android App Builder vLaunch (May 2026)

Google AI Studio can now build entire native Android apps from a single prompt — no installation required. Powered by Gemini, it generates Kotlin/Jetpack Compose apps with an embedded browser-based Android emulator for live preview, direct device install via USB/ADB, and one-click publish to Google Play internal testing track.

Google Colab CLI v1.0

The Google Colab Command-Line Interface (CLI) is a tool that bridges local terminals and remote Colab runtimes, enabling developers and AI agents to execute scripts, download models, and automate machine learning pipelines seamlessly. It offers features like instant GPU/TPU provisioning, remote execution of Python scripts, artifact recovery, and interactive access to remote environments.

Angular v22.1.1

Angular is a comprehensive, open-source development platform and framework for building scalable, high-performance web applications. It leverages TypeScript to provide a robust architecture for mobile and desktop web development, maintained by a dedicated team at Google.

Antigravity CLI v2.0 (May 2026)

Antigravity CLI is Google's terminal-first agentic coding tool — the successor to Gemini CLI — powered by Gemini 3.5 Flash. Launched at Google I/O 2026 as part of Antigravity 2.0, it lets developers orchestrate multi-agent workflows, schedule background tasks, and build custom agents from the terminal with full Google Cloud, Android, Firebase, and AI Studio integration.

Nano Banana 2 Lite v2 Lite

Nano Banana 2 Lite is Google's most efficient Gemini Image model, designed for rapid image generation and editing with minimal latency and cost. It offers high-quality outputs while maintaining the control and accuracy expected from the Nano Banana series.

Gemini CLI 0.33.0 v0.33.0

Gemini CLI is an open-source command-line interface developed by Google that integrates the power of Gemini AI models directly into your terminal. It enables developers to interact with Gemini models seamlessly, enhancing productivity and workflow efficiency. v0.33.0 introduces Plan Mode: Introducing a read-only environment for researching, designing and planning before implementing. A read-only mode that allows you to safely explore and work with Gemini CLI to come up with a plan prior to implementation. Plan mode can leverage read-only MCP tools, Agent Skills and is built to be fully extensible

Magika v1.1.0

Magika is an AI-powered file type detection tool that utilizes deep learning to provide fast and accurate identification of file contents. It is designed to process hundreds of billions of files weekly, supporting both binary and textual formats with approximately 99% accuracy.

Google Code Wiki v2026-05-08

Code Wiki is a new perspective on development for the agentic era. Gemini-generated documentation that's always up-to-date. Automatically generates and maintains interactive knowledge bases from code repositories, with AI agents that update docs when PRs merge, diagrams that visualize architecture, and natural language chat to understand your codebase.

WaxalNLP v1.0.0

The WaxalNLP dataset is a comprehensive collection of audio data for both Automated Speech Recognition (ASR) and Text-to-Speech (TTS) tasks in 14 African languages. It aims to enhance the accuracy and fluency of speech and language technologies for underserved African languages and serves as a resource for digital preservation.

Agentic Vision in Gemini 3 Flash vGemini 3 Flash

Agentic Vision is a new capability in Google's Gemini 3 Flash model that combines visual reasoning with code execution to ground answers in visual evidence. This feature enables the model to actively manipulate and analyze images, enhancing its ability to provide accurate and contextually relevant responses.

ADK for Go 2.0 v2.0

ADK for Go 2.0 is a framework for building complex, multi-agent Go applications using a graph-based workflow engine. It provides built-in support for human-in-the-loop interaction, dynamic orchestration, and state management within an idiomatic Go environment.

Google Antigravity SDK v0.1.6

The Google Antigravity SDK is a Python library that provides developers with programmatic access to the Google Antigravity agent harness. It allows for the creation, testing, and deployment of autonomous AI agents using the same core infrastructure and tools that power Google's Antigravity 2.0 and CLI products.

Google Workspace CLI

The Google Workspace CLI is a command-line tool developed by Google that enables users to manage various Google Workspace services, including Drive, Gmail, Calendar, Sheets, Docs, Chat, and Admin, directly from the terminal. It is dynamically built from Google's Discovery Service and incorporates AI agent capabilities to enhance user experience.

Universal Commerce Protocol (UCP) v1.0

UCP is an open-source standard developed by Google to facilitate seamless agentic commerce experiences. It enables AI agents and merchant systems to communicate effectively, allowing users to complete purchases within a single conversation. By providing a shared language, UCP eliminates the need for custom integrations between retailers and various platforms. ([marktechpost.com](https://www.marktechpost.com/2026/01/12/google-ai-releases-universal-commerce-protocol-ucp-an-open-source-standard-designed-to-power-the-next-generation-of-agentic-commerce/?utm_source=openai))

A2UI v0.8

A2UI is an open-source specification and set of libraries developed by Google that enables agents to describe rich native interfaces in a declarative JSON format. This approach allows client applications to render these interfaces using their own components, facilitating secure and interactive user interfaces across trust boundaries without the need to send executable code.

Gemini 3.1 Pro v3.1 Pro

Gemini 3.1 Pro is Google's latest AI model designed to tackle complex tasks requiring advanced reasoning. It offers improved performance over its predecessor, Gemini 3 Pro, with a verified score of 77.1% on the ARC-AGI-2 benchmark, more than doubling the reasoning performance of Gemini 3 Pro. ([blog.google](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/?utm_source=openai))

Gemini 3.6 Flash v3.6 Flash

Gemini 3.6 Flash is a high-performance, cost-efficient AI model optimized for coding, knowledge work, and complex multimodal tasks. It builds upon the capabilities of its predecessor, 3.5 Flash, by reducing token usage by up to 17% and streamlining multi-step reasoning workflows.

WebMCP vEarly Preview

WebMCP is an initiative by Google to standardize how AI agents interact with websites, aiming to enhance the efficiency and reliability of these interactions. By defining structured tools, WebMCP enables AI agents to perform actions on websites with increased speed and precision.

Google Antigravity 2.0 v2.0

Google Antigravity 2.0 is an agentic development platform designed to orchestrate multiple autonomous AI agents across complex software projects. It acts as a central command center for developers, enabling parallel task execution, background automation, and cross-surface integration.

Gemma 4 12B v4 12B

Gemma 4 12B is a unified, encoder-free multimodal model designed to deliver high-performance AI capabilities directly to laptops. It integrates vision and audio inputs seamlessly into its language model backbone, enabling advanced reasoning and agentic workflows without the need for separate encoders. This model is optimized for local deployment, requiring only 16GB of VRAM or unified memory, making it accessible for developers seeking powerful AI tools on standard hardware.

Gemini API Event-Driven Webhooks v2026-05-04

Google has launched event-driven webhook support in the Gemini API, enabling developers to build complex long-running agentic applications without polling. Instead of repeatedly checking job status, developers register a webhook URL and Gemini calls it automatically when long-running tasks complete — reducing latency, cutting infrastructure overhead, and making async AI pipelines significantly simpler to build.

Interactions API vGeneral Availability

Google's Interactions API is a unified interface for interacting with Gemini models and agents, enabling developers to build advanced applications with complex interactions, server-side state management, background execution, and multimodal generation.

AX (Agent Executor)

AX (Agent Executor) is Google's open-source distributed agent runtime for production AI agents. It coordinates agentic loops across distributed isolated actors (skills, tools, agents), manages durable execution state with automatic recovery and resumption, and provides a single-writer architecture for consistent state management — designed to scale from single machines to Kubernetes clusters.