#google deepmind
Google Deepmind: 29 AI articles covering google deepmind news, analysis, and research
Articles
Google Releases Gemini 4 Hydrogen: A 2.5x Faster Model forβ9
Google releases Gemini 4 Hydrogen, a fast 2.5x coding and agentic model with a 1M-token window, strong benchmarks, and closed API-only access at low cost.
Google DeepMind Staff Turn to UK Government as Unionizationβ9
Google DeepMind staff are escalating their unionization bid to the UK government after talks with management stalled, testing labor law at a leading AI lab.
Google DeepMind Releases AlphaGenome 2 for Cell-Type-Specificβ8
Google DeepMind releases AlphaGenome 2, a cell-type-specific DNA model that predicts variant effects, outperforming prior models on single-cell benchmarks via A...
Google DeepMind's New AI Tool Formally Verifies the Signalβ9
Google DeepMind's new AI tool formally verifies the Signal Protocol's security, using AI to mathematically prove the widely used encryption standard works as in...
Google Releases Gemini Omni 1.2 Flash: 15-Second AI Video at $0β8
Google releases Gemini Omni 1.2 Flash, generating natively 15-second AI video at $0.07 per second, with 30% lower pricing than Omni 1.1 Flash on Google Cloud an...
Google DeepMind Unveils Gemini 4 Carbon: A Frontier Model forβ8
Google DeepMind's Gemini 4 Carbon leads 15 of 17 benchmarks for agentic coding and cyber defense, but limited access keeps it from public developers for now.
Google DeepMind's Genie 4: A Real-Time World Model With Nativeβ8
Google DeepMind's Genie 4 world model generates real-time, explorable worlds with native audio and an in-loop Gemini planner. Here's what it means for agent tra...
Google's Genie 4 World Model Adds Audio, More Realistic Worlds,β9
Google DeepMind's Genie 4 world model adds real-time audio, photorealistic interactive worlds, mid-simulation prompt steering, and two-minute scene consistency.
Elorian's First AI Model Wants to Reason With Images, Not Justβ9
Elorian launches Glimpse, a research preview AI model that reasons with images via a visual scratch space, enabling crop, zoom, and frame-level inspection.
Google Releases Gemini 4 Argon API With 1M Token Output Limitβ8
Google releases Gemini 4 Argon API, a sparse MoE model with a 1M token output limit, $2/$10 pricing, and top scores in long video, knowledge work, and code.
