Conversation
Argon brings major improvements in spatial reasoning, mathematical precision, and visual coding.
Using multi-agent loops in , it autonomously built this entire pagoda complex, sculpting the architecture, foliage, and lighting from scratch in Blender.
The media could not be played.
In robotics, Gemini 4 Argon is speeding up real-world physical learning.
We let Argon watch a video of a human playing cornhole. It mapped the environment, calculated whole-body joint movements, and autonomously built the simulation code to practice.
When transferred to Apptronik’s Apollo humanoid, the robot sank seven straight shots in a real-world trial.
The media could not be played.
Post your reply
Gemini 4 Argon is available now for Google AI Ultra and Pro subscribers in and AI Mode in Search, Ultra subscribers in , and to developers and enterprises via the Gemini API in and the Gemini platform.
Argon will launch at an introductory
Scientists and researchers can apply for a chance to get access to ’s highest usage tier free of charge for 6 months, and cybersecurity defenders can apply via the Fairwind Program.
Gemini 4 Argon is just the beginning. We’re looking forward to bringing you the rest
The next generation Starlink V5 has a smaller form factor and lightweight design with greater power efficiency.
With speeds up to 375+ Mbps, Starlink V5 delivers reliable home internet for streaming, video calling, gaming and more.
Available in select areas.
x.com/xdotli/status/ best way to use gemini 4!
Quote
Xiangyi Li
@xdotli
Introducing Robo Use: API to physical worlds for agents
One interface, 32 embodiments, 100+ scenarios, and 11 agents and models
We aggregated all academic simulator scenarios and ran 2500+ trials with initial results.
Preview them or try it yourself 🧵
The media could not be played.
Translated from Chinese
Corntoss hits seven times in a row, with both stability and generalization ability being impressively stunning.
Awesome work, GDM team!!
I've a so101 robotic arm, wouldn't it be cool to make a create harness and test embodied AI with new Argon model. Only if I had ultra to do the job ;)
The video-to-sim-to-real pipeline is the interesting part here. Was Argon optimizing a reusable control policy, or synthesizing task-specific trajectories/programs through simulation?
