What DeepMind shipped in February 2026

Nano Banana 2: Combining Pro capabilities with lightning-fast speed

Our latest image generation model offers advanced world knowledge, production ready specs, subject consistency and more, all at Flash speed.

Gemini 3.1 Pro: A smarter model for your most complex tasks

3.1 Pro is designed for tasks where a simple answer isn’t enough.

Google has launched Gemini 3.1 Pro, an upgraded AI model designed to handle complex reasoning tasks that require more than simple answers. The model is rolling out across consumer and developer platfo

A new way to express yourself: Gemini can now create music

The Gemini app now features our most advanced music generation model Lyria 3, empowering anyone to make 30-second tracks using text or images.

Accelerating discovery in India through AI-powered science and education

Google DeepMind brings National Partnerships for AI initiative to India, scaling AI for science and education

Accessing WeatherNext on VertexStay organized with collectionsSave and categorize content based on your preferences.

Google introduced WeatherNext 2 on Vertex AI, enabling users to generate custom, on-demand weather forecasts instead of relying on static datasets. Access is limited to an Early Access Program (EAP) w…

Publications

The webpage is a collection of recent research publications by DeepMind on complex and interesting challenges in AI. It lists 238 publications, showcasing the breadth and depth of their work in the fi…

Simplicity and Complexity in Combinatorial Optimization

The article explores the connections between Kolmogorov complexity, optima, and optimization in combinatorial optimization problems, which are prevalent in physics and computer science. It argues that…

Run Gemma content generation and inferences

The article provides a guide on running Gemma models, focusing on the key decisions of choosing a Gemma variant and an AI execution framework. It highlights the importance of hardware availability and…

Gemini 3 Deep Think: Advancing science, research and engineering

Our most specialized reasoning mode is now updated to solve modern science, research and engineering challenges.

Accelerating Mathematical and Scientific Discovery with Gemini Deep Think

Research papers point to the growing impact of Deep Think across fields

Hybrid neural–cognitive models reveal how memory shapes human reward learning

A study challenges traditional reinforcement learning (RL) models by demonstrating that human reward learning requires flexible memory variables to track rich representations of the past. The research…