Google DeepMind
@GoogleDeepMind · view on X ↗
The engine room of @Google. Building AI safely and responsibly to solve the world’s most complex problems. Join us: https://t.co/jUHQA27iBL
- 1.5M followers
- 6 threads on ThreadOut
Unrolled threads
Every thread by Google DeepMind that readers unrolled on ThreadOut.
- May 19, 2026
We’re dropping Gemini Omni: our first step towards a model that can create anything from anything - starting with video. It combines Gemini’s intelligence with our generative media systems - represen...
SummaryGoogle announces Gemini Omni, a new AI model that combines advanced intelligence with generative media capabilities to create videos from text descriptions. The model maintains character consistency across scenes, understands physics and logic, and allows editing through natural language or style references.5 tweets1 min read - May 12, 2026
We’re reimagining a 50-year-old interface - the mouse pointer - with AI. 🖱️ These experimental demos show how people can intuitively direct Gemini on their screens using motion, speech, and natural s...
SummaryGoogle is experimenting with AI-enhanced mouse pointer technology that combines gestures and speech to provide contextual assistance. The system understands what's under your cursor, allowing natural interactions like pointing at a PDF to request bullet points or highlighting text to modify it.6 tweets1 min read - Apr 2, 2026
Meet Gemma 4: our new family of open models you can run on your own hardware. Built for advanced reasoning and agentic workflows, we’re releasing them under an Apache 2.0 license. Here’s what’s new...
SummaryGoogle releases Gemma 4, a new family of open-source models in four sizes (31B dense, 26B MoE, E4B, E2B) under Apache 2.0 license, designed for advanced reasoning, agentic workflows, and local deployment on custom hardware with up to 256K context window. The models support tool use, multi-step task execution, and mobile deployment with vision and audio capabilities.4 tweets1 min read - Mar 3, 2026
Gemini 3.1 Flash-Lite has landed. It’s our most cost-efficient Gemini 3 series model yet, built for intelligence at scale. Here’s what’s new 🧵
SummaryGoogle has released Gemini 3.1 Flash-Lite, a cost-efficient AI model that outperforms its predecessor with faster performance at lower costs. The model introduces adjustable 'thinking levels' for customized reasoning across different tasks and is now available in preview through the Gemini API.3 tweets1 min read - Nov 18, 2025
This is Gemini 3: our most intelligent model that helps you learn, build and plan anything. It comes with state-of-the-art reasoning capabilities, world-leading multimodal understanding, and enables...
SummaryGoogle announces Gemini 3, their most advanced AI model featuring state-of-the-art reasoning, multimodal capabilities, and improved coding abilities. Gemini 3 Pro is now rolling out globally and tops multiple AI benchmarks while offering more accurate, nuanced responses without filler.6 tweets1 min read - Nov 13, 2025
SIMA 2 is our most capable AI agent for virtual 3D worlds. 👾🌐 Powered by Gemini, it goes beyond following basic instructions to think, understand, and take actions in interactive environments – mea...
SummaryGoogle announced SIMA 2, an advanced AI agent powered by Gemini that can understand and act in 3D virtual worlds with improved reasoning, generalization to unseen environments, and self-teaching capabilities. The system represents progress toward more capable AI agents with potential applications in robotics and real-world AGI development.5 tweets1 min read