Blog

ICLR 2026 — hands-on pipeline from teleop collection through unified latent VLA training to whole-body loco-manipulation deployment on AgiBot X2.

A step-by-step guide to Ark v1.5 — Python-first robot learning framework with a Gym-style API, native ROS, ACT and Diffusion Policy out of the box.

Detailed guide to Booster Gym — open-source end-to-end RL framework training Booster T1 humanoid walking from teleop to real-world deployment.

NVIDIA NVlabs proves action as text reaches 94.7% on LIBERO, beating pi_0 and GR00T-N1 with zero architecture modification — just Qwen2.5-VL-3B.

Detailed guide to RDT2 from THU-ML — a foundation model that zero-shot deploys to bimanual UR5e and Franka, with open-source code.

Step-by-step guide to train SO-101 robot arm in NVIDIA Isaac Lab, collect teleop data, fine-tune GR00T N1.5, and deploy the policy on real hardware.

Beginner-friendly walkthrough of Wheeled Lab — open-source ecosystem to train RC cars to drift, climb, and visually navigate in Isaac Lab and deploy zero-shot t

Reproduce LIFT (ICLR 2026): pretrain SAC in MuJoCo Playground, then sim2sim fine-tune a physics-informed world model in Brax for Booster T1 and Unitree G1.

How Google DeepMind built a table tennis robot that beats amateur humans using hierarchical RL, LLC-HLC architecture, and zero-shot sim-to-real transfer.

Step-by-step guide to install and run Xiaomi-Robotics-0 — a 4.7B VLA combining Qwen3-VL and Diffusion Transformer with 80ms real-time inference on RTX 4090.

Explore Dense Object Nets, DenseFusion, and DenseMatcher — three dense visual descriptor technologies revolutionizing how robots see and grasp objects.

OpenHelix: the open-source dual-system VLA pairing LLaVA-7B (System 2) with a 3D diffusion policy (System 1) via a learned ACT token, achieving SOTA on CALVIN.