Yesterday, we live demo-ed a “generalist” VLA for (I think) the first time ever to a broad audience @RoboticsSciSys. Bring any object. Ask anything. New environment, new instructions, no fine-tuning. Just impeccable vibes! ✨.

RT @ayzwah: We took a robot to RSS in LA running our new Gemini Robotics On-Device VLA model. People interacted with the model with new obj….

Thanks for cheering @GautamSalhotra @Ishika_S_ @RussTedrake @_abraranwar and our favorite Prof. Rudy who is still not on Twitter!.

This is the Gemini Robotics On-Device VLA that runs on a single GPU: and you can apply for access today!. Shoutout to @ayzwah @debidatta @SudeepDasari @ashwinb96 @xjygr08 @xiao_ted @sippeyxp @jackyliang42 @TonyWentaoYuan @ColinearDevin and everyone else at GDM Robotics for making.

RT @KuanFang: Join us for a full day of exciting talks and discussions on learning representations for robotic intelligence!.Learned Robot….

Huh. @grok is this true?.

Busy day with Gemini Robotics On-Device on Aloha! We’re here all afternoon, come find us.

Come by the @GoogleDeepMind booth at @RoboticsSciSys conference in LA!. We’re demoing Gemini Robotics On-Device live, come check it out

Oh and it works on Aloha, Frankas, and the @Apptronik humanoid!. open-source SDK:. open-source sim:

It's been a wonderful journey working on this and am really excited to see what the growing community of researchers and builders will make with this model. We will be doing *live demos all day* @RoboticsSciSys today in LA -- stop by our booth to interact with the model and test.

Excited to release Gemini Robotics On-Device and bunch of goodies today.🍬 on-device VLA that you can run on a GPU.🍬 open-source MuJoCo sim (& benchmark) for bimanual dexterity.🍬 broadening access to these models to academics and developers.

RT @GoogleDeepMind: We’re bringing powerful AI directly onto robots with Gemini Robotics On-Device. 🤖. It’s our first vision-language-actio….

In Nashville for my first #CVPR! 🎸. I'll be speaking about evaluating and improving VLAs at two workshops:. Thurs 10a: Robotics 101 Thurs 2.30p: Real2Sim Hit me up to chat robotics, "embodied AI", grounded reasoning and more!.

RT @smch_1127: 🚀 HERE WE GO! Join us at CVPR 2025 for a full-day tutorial: “Robotics 101: An Odyssey from a Vision Perspective”.🗓️ June 12….

Gemini Robotics makes first contact with the outside world at #GoogleIO this week!. “this is the first time in my life that I’ve been able to control robots using nothing but my voice. That is basically the very definition of cool.” @AndroidAuth. “Gemini Robotics is exactly the

Excited to be in Atlanta for #ICRA2025! 🍑🦾. We’ll be presenting some exciting research on evaluating and improving generalization for robot policies. 📋 Thurs AM: improving STEERability via relabeling w/ @smithlaura1028. 🎤 Fri: speaking at the SocialNav workshop. Come say hi!.

FSC is an effective way to guide data collection starting from scratch (e.g. Diffusion Policy) to SoTA VLAs (π0) — in both real & sim!. Open-source code coming soon!. Wonderful collab led by @LihanZha @ApurvaBadithela @allenzren @Majumdar_Ani & others. See Lihan's 🧵⤵️ for.

🗝️ Idea:.Start with a nominal dataset → vary one factor at a time → fit a power-law scaling curve to performance (real or proxy like embedding distance) → pick factor with max slope to collect next. Efficiently improve performance with just O(100) demos!

In the era of generalist robot foundation models, how do you get their pre-trained model to work well on your robot and task?. 🌐 📈 We introduce Factored Scaling Curves (FSC): a principled approach for modeling how policy performance scales with data for

We trained an end-to-end visual navigation policy on 2000 hours of uncurated, crowd-sourced data and evaluated it across 24 environments in 6 countries! 🌎🤖. Recipe: Train a model-based policy on a small amount of clean short-horizon data --> use it to re-label actions for the