IDENTITY / 01Image source · Google DeepMind · Public domain · Monochrome treatment; original available.
LAB
Google DeepMind
Connected to 73 published stories. Open a story for context and original sources.
Google introduces Gemini 3.8 text-to-speech models
Read story →Google releases AlphaGenome Atlas predictions for nine billion DNA variants
Read story →Google begins rolling out Gemini Omni 1.1 Flash for video
Read story →Google introduces Gemini 3.7 Flash and adds it to Gemini Spark
Read story →Google introduces Gemini Robotics 2 for whole-body control
Read story →Google introduces Gemini 3.5 Flash Cyber for vulnerability work
Read story →Google releases Nano Banana 2 Lite and opens the Omni Flash API
Read story →Google offers Co-Scientist hypothesis generation to researchers
Read story →Google introduces Gemini Omni Flash for video generation
Read story →Google releases Gemini 3.5 Flash and tests the Gemini Spark agent
Read story →Google reports AlphaEvolve gains in genomics and Earth modeling
Read story →Google reports more resilient distributed training with Decoupled DiLoCo
Read story →Google releases Gemini Robotics-ER 1.6 through its API
Read story →Google releases Gemma 4 model weights under Apache 2.0
Read story →Google rolls out Lyria 3 Pro for longer music generation
Read story →Google releases Gemini 3.1 Pro across developer and consumer tools
Read story →Google opens Project Genie to US AI Ultra subscribers
Read story →Google launches Gemini 3 in its app and Search
Read story →Google DeepMind previews Genie 3 interactive world generation
Read story →Google reports Gemini Deep Think reached gold-medal standard on IMO problems
Read story →Google introduces Veo 3 with generated audio and the Flow filmmaking tool
Read story →Google DeepMind introduces AlphaEvolve for algorithm discovery
Read story →Google previews Gemini 2.5 Pro with reasoning before responses
Read story →Google DeepMind introduces Gemini Robotics and Robotics-ER
Read story →Google introduces experimental Gemini 2.0 Flash and agent prototypes
Read story →Google DeepMind releases Gemma Scope tools for studying Gemma 2
Read story →Google DeepMind and Isomorphic Labs introduce AlphaFold 3 for molecular complexes
Read story →Google previews Gemini 1.5 Pro with a million-token context window
Read story →Google DeepMind combines neural suggestions and symbolic proofs in AlphaGeometry
Read story →Google introduces Gemini and starts rolling out Pro and Nano
Read story →Google DeepMind reports 2.2 million predicted crystal structures
Read story →Google DeepMind trains RT-2 to turn visual-language knowledge into robot actions
Read story →Google combines DeepMind and Google Brain into Google DeepMind
Read story →DeepMind uses AlphaTensor to discover matrix-multiplication algorithms
Read story →DeepMind and EMBL-EBI expand AlphaFold to over 200 million predictions
Read story →Nando de Freitas argues that scaling could deliver AGI after Gato
Read story →DeepMind trains Gato on 604 tasks with one set of model weights
Read story →DeepMind introduces Flamingo for few-shot image and video tasks
Read story →DeepMind finds language models need more training data for a fixed compute budget
Read story →DeepMind and EPFL control tokamak plasma with reinforcement learning
Read story →DeepMind introduces AlphaCode for competitive programming
Read story →DeepMind evaluates 280-billion-parameter Gopher across 152 tasks
Read story →DeepMind’s RETRO retrieves text from a two-trillion-token database
Read story →DeepMind trains XLand agents across changing multiplayer tasks
Read story →DeepMind and EMBL-EBI open an AlphaFold structure database
Read story →DeepMind publishes AlphaFold2’s method and releases prediction code
Read story →AlphaFold achieves high accuracy in the CASP14 protein-structure assessment
Read story →Google Maps uses graph networks to improve arrival-time estimates
Read story →Agent57 exceeds the standard human baseline across 57 Atari games
Read story →Researchers evaluate breast-screening AI on UK and US mammograms
Read story →AlphaStar reaches Grandmaster with all three StarCraft races
Read story →DeepMind and VA study prediction of future acute kidney injury
Read story →DeepMind introduces AlphaStar after wins over StarCraft professionals
Read story →DeepMind publishes AlphaZero’s full evaluation in Science
Read story →DeepMind’s AlphaFold places first at CASP13
Read story →DeepMind reports direct AI control of data-centre cooling
Read story →DeepMind and Moorfields report AI referral recommendations from retinal scans
Read story →DeepMind introduces a network that renders unseen views of simulated scenes
Read story →DeepMind and Android announce adaptive battery and brightness features
Read story →DeepMind appoints Lila Ibrahim as its first chief operating officer
Read story →Google opens Cloud Text-to-Speech beta with WaveNet voices
Read story →DeepMind introduces IMPALA and releases DMLab-30
Read story →DeepMind reports AlphaZero results in chess, shogi and Go
Read story →DeepMind introduces AlphaGo Zero trained through self-play
Read story →AlphaGo completes a 3–0 match victory over Ke Jie
Read story →DeepMind researchers introduce the Kinetics action-video dataset
Read story →Technology companies establish the Partnership on AI
Read story →DeepMind introduces WaveNet for generating raw audio
Read story →DeepMind reports up to 40% less energy for data-center cooling
Read story →AlphaGo defeats Lee Sedol 4–1 in Seoul
Read story →Mnih and colleagues introduce asynchronous reinforcement learning
Read story →DeepMind publishes AlphaGo’s 5–0 win over Fan Hui
Read story →DeepMind trains AI to play 49 Atari games from pixels and scores
Read story →