Ce que DeepMind a publié en mars 2026

Reimagining the mouse pointer for the AI era

Google DeepMind is transforming the mouse pointer into a context-aware AI partner. Move beyond the friction of traditional prompting with intuitive AI collaboration in Chrome and beyond.

Gemini 3.1 Flash Live: Making audio AI more natural and reliable

DeepMind has introduced a new voice model, Gemini 3.1 Flash Live, designed to enhance audio AI capabilities by improving precision and reducing latency. This update aims to make voice interactions mor…

Protecting people from harmful manipulation

Google DeepMind has published research highlighting risks of harmful manipulation posed by AI systems, particularly in critical domains such as finance and healthcare. The company introduces new safet…

Lyria 3 Pro: Create longer tracks in more

DeepMind has introduced Lyria 3 Pro, a significant upgrade to its music generation model, enabling the creation of longer tracks with improved structural awareness. This advancement allows for more co…

Strategic Tradeoffs Between Humans and AI in Multi-Agent Bargaining

Measuring progress toward AGI: A cognitive framework

DeepMind has introduced a new cognitive framework designed to measure progress toward Artificial General Intelligence (AGI). This framework aims to provide a standardized method for evaluating advance…

The Abstraction Fallacy: Why AI Can Simulate But Not Instantiate Consciousness

The article challenges the dominant view of computational functionalism in AI consciousness debates, arguing that subjective experience does not emerge from abstract causal topology. It introduces the…

From games to biology and beyond: 10 years of AlphaGo’s impact

Ten years after the launch of AlphaGo, DeepMind reflects on its impact and how it is driving scientific discovery. AlphaGo, a significant milestone in AI development, has paved the way for advancement…

Our most accurate AI weather forecasting technology

Google DeepMind and Google Research have launched WeatherNext 2, a state-of-the-art family of AI-powered weather forecasting models. This update is significant as it provides faster, more accurate, an…

Nano Banana 2 🍌

Nano Banana 2 is a state-of-the-art, Pro-level image generation and editing tool that offers flash-level speed. It leverages the Gemini model’s real-world knowledge base and real-time information from…

Lyria 3

DeepMind has launched Lyria 3, its most advanced AI music generation model to date. This tool allows users to create high-fidelity music by uploading images, defining vocal styles, and exporting profe…

Genie 3

DeepMind has introduced Genie 3, a general-purpose world model that represents a significant advancement in AI capabilities. Genie 3 can generate photorealistic environments in real-time from simple t…

Veo

DeepMind has introduced Veo 3.1, a state-of-the-art video generation model designed to empower filmmakers and storytellers. This latest model offers greater control, consistency, and creativity, with …

Gemini 3

Google DeepMind has introduced Gemini 3, its most intelligent AI model yet, which brings significant advancements in reasoning, multimodal understanding, and agentic capabilities. This new model is de…

Gemini Audio

Google's DeepMind has introduced Gemini Audio, a suite of advanced real-time audio models built on the Gemini platform. This new offering includes live audio agents for natural, real-time conversation…

Nano Banana 🍌

Google DeepMind has introduced new image generation and editing models under the Gemini brand, including Nano Banana, Nano Banana 2, and Nano Banana Pro. These models leverage advanced AI capabilities…

Imagen

The webpage introduces Imagen 4, the latest iteration of Google DeepMind's text-to-image model, highlighting its enhanced capabilities and improvements. Imagen 4 is designed for creativity, offering p…

Vertex AI Studio

Vertex AI Studio Explore 200+ models on our enterprise platform with tools and features for AI development Try Vertex AI Studio

Google Antigravity

Google Antigravity Experience an agentic development platform, evolving the IDE into the agent-first era Try Google Antigravity

Google AI Studio

Google AI Studio Start building something new, with cutting-edge AI models and tools Try Google AI Studio

Gemma

Gemma Our family of state-of-the-art, open models Build with Gemma Run Gemma Developer docs Learn more

Gemini Audio Advanced real-time audio models, built on Gemini Try it in Google AI Studio Vertex AI Studio Learn more

Lyria

New Lyria Our most advanced music generation model yet Try it in Gemini app Learn more

Gemini

Gemini Our most intelligent AI models Try it in Gemini app Google AI Studio Google Antigravity Vertex AI Studio Learn more

Weather Lab

Weather Lab Test our experimental weather forecasting models. Learn more

Gemini 3.1 Flash-Lite: Built for intelligence at scale

Gemini 3.1 Flash-Lite is our fastest and most cost-efficient Gemini 3 series model yet.

Gemini 3.1 Flash-Lite: Built for intelligence at scale March 2026 Models Learn more