Ce que DeepMind a publié en 2026

EmbeddingGemma 2: an open, lightweight multimodal embedding model

DeepMind introduced EmbeddingGemma 2, a new open-source, lightweight multimodal embedding model designed for efficiency and versatility. The model supports both text and image inputs, enabling develop…

Gemini 4 Argon: our next era of frontier intelligence

DeepMind introduced Gemini 4 Argon, positioning it as the next generation of frontier AI models. The announcement emphasizes breakthroughs in reasoning, multimodal capabilities, and scalability, marki…

Introducing Gemini 3.8 Live with Live Avatar

DeepMind introduced Gemini 3.8 Live, a new version of its AI model featuring a live avatar capability that enables real-time, interactive 3D avatars for user engagement. The feature is designed to enh…

Advancing Private AI Compute with secure, server-side memory

DeepMind introduced private, server-side memory to its Private AI Compute platform, enabling personal AI systems to retain secure context without storing data locally. This enhancement aims to improve…

Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking

DeepMind introduced Gemini 3.8 Live and its Extended Thinking variant, enhancing real-time reasoning and problem-solving. The models are designed for live interactions, offering improved contextual un…

AlphaGenome — Google DeepMind

Google DeepMind introduced AlphaGenome Atlas, a 1PB dataset predicting the molecular impact of all nine billion single-letter human genome variants using the AlphaGenome Variant Impact (AVI) score. In…

AlphaGenome Atlas: A predictive map of every possible DNA letter change in the human genome

DeepMind introduced AlphaGenome Atlas, a predictive map detailing every possible DNA letter change in the human genome. This tool aims to accelerate genetic research by providing a comprehensive refer…

Gemini Audio

New Gemini Audio Advanced real-time audio models, built on Gemini Try it in Google AI Studio Gemini Enterprise Agent Platform Learn more

Gemini 3.5 Transcribe

Gemini 3.5 Transcribe Our most precise speech-to-text model yet, designed for intelligent voice interactions. Learn more Try

Gemini Omni 1.1 Flash

Your browser does not support the video tag. Your browser does not support the video tag. Gemini Omni 1.1 Flash Greater creative control and production-ready outputs Learn more Try

Gemini 3.8 Flash

Your browser does not support the video tag. Your browser does not support the video tag. Gemini 3.8 Flash Best for tackling complex agentic tasks at scale Learn more Try

Introducing WeatherNext 3, our most advanced and accurate global weather AI model

Google DeepMind introduced WeatherNext 3, a significantly improved global weather AI model claimed to be the most accurate to date. The model leverages advanced machine learning techniques to enhance …

Google Flow - AI Creative Studio for Video, Images & Custom Tools

Google introduced Flow, an AI-powered creative studio leveraging its generative models (Gemini Omni, Nano Banana, Veo 3.1) for video editing, image generation, and custom tool creation. The platform i…

Google DeepMind introduced Gemini 3.8 Flash, positioning it as their most capable workhorse model optimized for agentic tasks, coding, and long-horizon workflows. The model delivers advanced reasoning

Introducing Gemini 3.8 Flash and 3.8 Flash Cyber

Google introduced two new variants of its Gemini 3.8 model: 3.8 Flash and 3.8 Flash Cyber. The 3.8 Flash Cyber is optimized for cybersecurity use cases, while 3.8 Flash serves as a general-purpose, hi…

Introducing agentic video understanding with Gemini

Google introduced agentic video understanding capabilities in its Gemini models, enabling advanced analysis and reasoning over video content. This feature allows the models to process, interpret, and …

Gemini Omni 1.1 Flash lets you build with more control

DeepMind announced the release of Gemini Omni 1.1 Flash, a version of their model offering more control for builders. The update emphasizes flexibility and customization in application development.

Introducing Gemini 3.7 Flash

Introducing Gemini 3.7 Flash

Gemini Robotics 2

Your browser does not support the video tag. Your browser does not support the video tag. Gemini Robotics 2 The intelligence layer to power any kind of robot Learn more Try

Gemini Robotics 2 brings whole body intelligence to robots

Gemini Robotics 2 brings whole body intelligence to robots