/
#57
Google DeepMind
Google's AI research lab behind the Gemini model family (2.5/3, with reasoning and multimodality), open-weight Gemma models (2B, 31B, MoE and dense variants), and AlphaFold for protein structure prediction. Also develops AlphaProof for formal RL-based theorem proving, Veo for video generation, and Gemini Robotics. Research spans reinforcement learning, interpretability, and AI safety.
Links