Gemini Robotics 1.5
Gemini Robotics 1.5 is a Vision-Language-Action (VLA) AI model by Google DeepMind — a software platform, not a physical robot. No physical dimensions (height, weight, DOF) apply. It runs on partner robot hardware including Apptronik Apollo humanoid, ALOHA 2, and Franka bi-arm systems. The model achieves 250ms observation-to-action latency and 50Hz control frequency. Physical specs belong to the partner robots it controls, not to Gemini Robotics 1.5 itself.
Primary use case: Enabling general-purpose robot control for complex multi-step physical tasks such as waste sorting, laundry organization, and dexterous manipulation across multiple robot form factors
Related news
Agility is public: the short list of humanoid robot stocks, and who's next
Physical Intelligence wants a $20 billion valuation. Here's the bull and bear case.
NVIDIA's massive robotics strategy: how all the puzzle pieces fit together
Google one-shots humanoid training, Tesla gives robots skin, Faraday launches $2,490 humanoids, and more!
One-shot robot training: Google's Gemini 4 Argon taught Apptronik's Apollo to play cornhole
China's humanoid robot lead widens, with domestic vendors supplying over 95% of global shipments
China's humanoid robot lead widens, with domestic vendors supplying over 95% of global shipments
China humanoid robot standards plan targets 100 key standards by 2028
China humanoid robot standards plan targets 100 key standards by 2028
DeepSeek, Unitree Join Forces to Develop Robotics AI Models
Chinese AI startup DeepSeek is investing 140.8 million yuan (approximately $19.6 million) in robotics manufacturer Unitree for a 2.31% stake, partnering to develop AI models for humanoid robots. Unitree is simultaneously proceeding with an IPO on Shanghai's STAR Market at a valuation of 61 billion yuan (~$8.5 billion), planning to raise approximately 6.1 billion yuan through the issuance of 40.45 million new shares. The partnership combines DeepSeek's AI model development capabilities with Unitree's mechanical engineering expertise and physical-world data to improve competitiveness in the humanoid robotics market against rival AgiBot.