Google DeepMind 推出三款實體 AI 模型：實現全身控制、靈巧操作與多機器人協作

Google DeepMind has released Gemini Robotics 2, the intelligence layer for its next generation of robots. The release moves the stack past table-top manipulation into whole body control, five finger dexterity and multi robot teamwork. It ships as three separate models with three different access tiers. Most robots today are pre-programmed or tele-operated for narrow, repetitive task sequences. They do not adapt to unpredictable environments, and skills rarely transfer between robot bodies. Gemini Robotics 2 targets all three limits at once. TL;DR Three models ship together: a VLA, an embodied reasoning VLM, and an on-device VLA. One checkpoint drives Apollo 2 with two different hands plus a Franka Duo gripper. Gemini Robotics ER 2 is public preview; the VLA and on-device models stay gated. Multi-finger dexterity is still the weak axis, ranging from 32% to 92%. ASIMOV-Agentic, a new safety benchmark, is on Hugging Face under CC-BY-4.0. 3 Models and What They Do Gemini Robotics 2 Gemini Robotics 2 is the vision-language-action (VLA) model. It converts vision and language input into motor control. It can drive full humanoids from feet to fingertips, and other bi-arm robots. It also handles dexterous manipulation on both multi-finger hands and parallel grippers. Gemini Robotics ER 2 Gemini Robotics ER 2 is the embodied reasoning (ER) model. It is a vision language model that acts as the high level brain. It communicates with humans, understands the physical world, and plans multi-step tasks lasting several minutes. Per its model card, ER 2 is based on Gemini 3.5 Flash. It accepts interleaved text, image, video and audio with a context window of up to 128k, and emits text up to 64K tokens. Gemini Robotics On-Device 2 Gemini Robotics On-Device 2 is the efficient VLA optimized to run locally on the robot. Its model card states it is built on Gemini Robotics 1.5 technology and Google’s on-device Gemma models. Inputs are text, images and robot proprioception as numerical values. Outputs are robot actions as numerical values. The division of labor matters for system design. ER 2 plans and tracks the task, then hands motor execution to a VLA declared as a tool. Developers register low level control interfaces, such as VLA models or navigation APIs, as callable tools. They then stream multimodal video, audio or text directly into the model. Whole body control on Apptronik Apollo 2 Previous Gemini Robotics models controlled only the humanoid’s upper body for table-top tasks. Gemini Robotics 2 extends control to whole body motion for the first time. The worked example uses Apptronik’s Apollo 2. Given the instruction “put the watering can into the green bin in the bottom shelf,” Apollo walks to the table and picks up the watering can. It then takes a few steps to the shelves and places the object at its destination. Google DeepMind is direct about the remaining gap. It states that its robots have more to advance in movement speed. Dexterity across hands and grippers Gemini Robotics 2 can control the five-fingered, 22 degree-of-freedom SharpaWave hand on Apollo 2. Reported actions include tying knots and sealing a ziplock bag. The same model also operates standard two-fingered parallel grippers on a Franka Duo platform for tasks like tight packing. The reported success rates come from one model checkpoint controlling three embodiments: Apollo 2 with SharpaWave hands, Apollo 2 with Inspire hands, and Franka Duo with a Robotiq gripper. CategoryEmbodimentTaskSuccessGeneral whole body manipulationApollo 2 + Inspire handsPick up from shelf76.3%General whole body manipulationApollo 2 + Inspire handsPick up from table68.4%General whole body manipulationApollo 2 + Inspire handsPick up from floor45.7%Multi-finger dexterityApollo 2 + Sharpa handsUnscrew bulb92%Multi-finger dexterityApollo 2 + Sharpa handsTie trash bag44%Multi-finger dexterityApollo 2 + Sharpa handsZiplock40%Multi-finger dexterityApollo 2 + Sharpa handsScrew bulb36%Multi-finger dexterityApollo 2 + Sharpa handsDustpan32%Gripper dexterityFranka DuoPrecise insertion tasks89.6%Gripper dexterityFranka DuoDiverse tool kitting78.9%Gripper dexterityFranka DuoGeneral pick and place74.2% ER 2: Temporal Intelligence and Tool Orchestration The developer X post focuses on a problem that rarely gets benchmarked: knowing when a task is actually done. Progress classification: Each frame in a video feed is assigned to one of five progress levels, from 0-20% through 80-100%. Gemini Robotics ER 2 reaches 57.4% accuracy on this task. Google DeepMind reports this outperforms previous generation models and competing frontier models. Moment finding: This measures whether the model can identify the exact frame where a critical event occurs. An example is the moment to stop pouring coffee into a cup. ER 2 reaches 91.3% accuracy with a 0.96 second mean absolute distance. Google DeepMind reports it competing closely with much larger model categories at 4x the execution speed. Tool orchestration: ER 2 is evaluated across three control modes: real VLA, sim VLA, and human tele-op. It outperforms Gemini Robotics ER 1.6 on all three. ER 2 integrates with the Gemini Live API through a bidirectional streaming endpoint. The stated purpose is removing the stop-and-think pauses that break multi-step execution. It can also natively call tools like Google Search or any user-defined function. Three spatial capabilities were upgraded. Success and failure detection now runs on raw video feeds instead of static snapshots, which catches mid-execution spills and slips. General instrument reading extends beyond circular dials to digital displays, linear scales, rulers and liquid thermometers, tested across 10 instrument types. Spatial visual question answering improves through Gemini’s multimodal advances. Google DeepMind built a demo with Spot from Boston Dynamics, using ER 2 to orchestrate Spot navigation and manipulator APIs. Sample code is available in the robotics-samples repository. Multi-Robot Collaboration Gemini Robotics 2 introduces collaboration between different robot types. The reasoning is that no single robot fits every task. A wheeled rover suits indoor work, while a humanoid handles uneven terrain better. Robots communicate through a shared semantic understanding to hand off subtasks. The demonstrated pairing is Apptronik’s Apollo 2 with a Franka F3 Duo. On-Device 2: Adapting to New Robot Bodies Gemini Robotics On-Device 2 targets applications that cannot rely on network latency or connectivity. It is natively multi-embodiment and inherits the motion transfer techniques from Gemini Robotics 1.5. Google DeepMind reports adaptation to new bi-arm embodiments in a few hours, typically with fewer than 200 examples. This holds for embodiments with drastically different shapes, sensors and degrees of freedom. Demonstrated platforms include Dexmate, SO101 and Trossen. The model card publishes a data scaling comparison against On-Device 1 on platforms introduced only during post-training: PlatformModelStartEndSO101On-Device 26.7%53.3%SO101On-Device 10.0%6.7%DexmateOn-Device 224.4%75.6%DexmateOn-Device 113.3%33.3% The SO101 result is the sharper one. On-Device 1 barely leaves the floor, while On-Device 2 crosses 50%. The model card also lists limits. On-Device 2 is limited in generalizing to out-of-distribution tasks and in controlling high-degree-of-freedom robots. Availability ModelAccessGemini Robotics ER 2Public preview via Gemini API and Google AI Studio; private preview on Gemini Enterprise Agent PlatformGemini Robotics 2 (VLA)Early-access partnersGemini Robotics On-Device 2Trusted Testers The AI Studio launch link uses the model string gemini-robotics-er-2-preview. Getting-started notebooks are in the robotics-samples repository. Interactive explainer (function(){ window.addEventListener("message",function(e){ if(e.data&&e.data.gr2height){ var f=document.getElementById("gr2frame"); if(f)f.style.height=e.data.gr2height+"px"; } }); })(); Sou