Google just rolled out Project Genie, an experimental research prototype powered by its Genie 3 world model, Nano Banana Pro, and Gemini. Instead of static 3D environments, it simulates physics and renders paths ahead in real time as you walk, fly, or drive through user-prompted worlds. For now, generations are capped at 60 seconds and limited to Google AI Ultra subscribers in the U.S.
Google DeepMind just debuted 'Dear Upstairs Neighbors' at the Sundance Film Festival, an animated short co-created with animation veterans. Instead of relying on text prompts, researchers built custom video-to-video workflows and local refinement tools powered by Veo and Imagen. Why it matters: It shows a practical path where professional animators keep precise creative control instead of wrestling with AI randomness.
Google DeepMind just introduced D4RT, a model that reconstructs and tracks dynamic 3D scenes directly from standard 2D video. By turning 4D perception into parallel spatial queries, it achieves up to 300x higher trajectory throughput than previous methods. Why it matters: it gives robots and AR glasses a path to real-time spatial intuition in moving environments.
Google DeepMind just open-sourced Gemma Scope 2, an interpretability suite designed to reverse-engineer how the Gemma 3 family actually thinks. Spanning models from 270M to 27B parameters, it uses sparse autoencoders and transcoders across every layer to map internal computations. Why it matters: It gives the safety community an open, microscopic look into why models hallucinate, get jailbroken, or disguise their reasoning.
Google DeepMind just made its feature-length documentary, “The Thinking Game,” free to stream on YouTube to mark AlphaFold's fifth anniversary. Filmed over five years by the crew behind AlphaGo, it follows founder Demis Hassabis and his team as they pursue AGI and crack a 50-year-old biology challenge. It offers a rare, intimate look at the human side of major scientific breakthroughs.
Google just launched Nano Banana Pro, an upgraded image generation and editing model built on Gemini 3 Pro. It can pull real-time search data to build accurate infographics, render crisp multilingual text in full paragraphs, and blend up to 14 reference images while keeping faces consistent. Free users can test it with limited quotas inside the Gemini app.
Google just added an AI image verification feature directly into the Gemini app. Users can now upload a picture and simply ask if it was AI-generated, prompting Gemini to inspect its imperceptible SynthID watermark and explain the context. Google is also embedding C2PA metadata into images created by its Nano Banana Pro model starting this week.
Google DeepMind just announced a new AI research lab in Singapore to accelerate AI adoption across the Asia-Pacific region. The team of scientists and engineers will focus on advancing Gemini with a strong emphasis on linguistic and cultural inclusivity. With DeepMind's APAC headcount already doubling over the past year, this marks a big push to adapt frontier models to local markets.
Google just unveiled Gemini 3, its next-generation frontier model family led by Gemini 3 Pro. The model claimed the top spot on LMArena with an unprecedented 1501 Elo, delivering huge upgrades in multimodal reasoning and vibe coding alongside a new agentic platform called Google Antigravity. Why it matters: Google is pushing hard to retake the AI crown by turning models into autonomous, self-executing software agents.
Google DeepMind just launched WeatherNext 2, a forecasting AI that runs 8x faster and generates hundreds of hourly weather scenarios in under a minute on a single TPU. It beats its predecessor across 99.9% of variables and is already rolling out across Google Search, Gemini, Pixel Weather, and BigQuery.
Google DeepMind just introduced SIMA 2, an AI agent powered by Gemini that plays and reasons inside 3D virtual worlds. Unlike its predecessor, it doesn't just follow simple commands; it operates via mouse and keyboard, explains its goals in natural language, and can self-improve in newly created Genie environments. It is currently available as a limited research preview for select academics and game developers.
Anthropic just revealed how it trains and measures political neutrality in Claude, open-sourcing a new automated benchmark. In paired prompt tests across hundreds of political stances, Claude Sonnet 4.5 scored on par with Gemini 2.5 Pro and Grok 4 while outperforming GPT-5 and Llama 4. Why it matters: Anthropic wants a shared industry standard so AI models avoid subtly nudging users toward partisan views.
Google just introduced Google Skills, a unified hub gathering nearly 3,000 courses, hands-on labs, and credentials from DeepMind, Cloud, and Grow with Google. It spans fast 10-minute tutorials to advanced research tracks, offering developers 35 free monthly lab credits alongside direct pipelines to a 150-employer hiring consortium.
OpenAI researchers just walked back claims that GPT-5 solved 10 decades-old Erdős math problems after experts pointed out the model merely surfaced existing papers that a database author had missed. DeepMind CEO Demis Hassabis called the sloppy communication "embarrassing," exposing the awkward gap between AI research hype and actual scientific progress.
Google DeepMind just partnered with Commonwealth Fusion Systems (CFS) to accelerate the path to net fusion energy. The collaboration combines DeepMind's plasma simulator TORAX with reinforcement learning to run millions of virtual simulations for CFS's compact SPARC tokamak, optimizing fuel injection and magnetic controls before turning the machine on. AI will also act as a real-time pilot to safely manage intense heat. Why it matters: Mastering plasma control could dramatically shorten the timeline to abundant, zero-carbon grid power.
Google just launched the Gemini 2.5 Computer Use model in public preview, letting developers build agents that navigate web and mobile interfaces by clicking, scrolling, and filling out forms. Built on Gemini 2.5 Pro, it tops browser benchmarks with lower latency and step-by-step safety controls. Why it matters: AI can now tackle messy real-world web chores that lack structured APIs.
Google DeepMind just unveiled CodeMender, an autonomous AI agent powered by Gemini Deep Think models that diagnoses and patches software vulnerabilities. Over six months, it has already upstreamed 72 security fixes to open-source projects, though human researchers still review every patch before submission.
Nvidia just launched Newton, an open-source GPU-accelerated physics engine co-developed with Google DeepMind and Disney Research, alongside its new Isaac GR00T N1.6 humanoid foundation model. Newton simulates messy, complex physical interactions like walking on snow or handling fragile objects. It gives developers a unified stack to train physical AI in simulation and deploy it safely to real robots.
Google DeepMind just unveiled Gemini Robotics 1.5 and Gemini Robotics-ER 1.5 to power physical AI agents. The ER model handles multi-step spatial reasoning and tool calling, while the 1.5 VLA model thinks before translating instructions into physical motor actions across different robot bodies. Developers can access the ER model today via the Gemini API in Google AI Studio.
More than 200 former heads of state, Nobel laureates, and top AI scientists just backed a global initiative demanding governments establish strict red lines for AI by the end of 2026. Proposed red lines include banning AI self-replication and human impersonation, aiming to curb catastrophic risks before an irreversible disaster strikes.
