Astribot releases SmoothRL for online learning on S1 - 2026-09-04 - Research Publication - SmoothRL separates committed, executed and discarded actions so policy gradients use only actions that physically ran. After 250 episodes on S1, reported success reached 94% for dynamic tossing, 83% for pen capping and 90%… - source: astribot.com
Astribot publishes Lumo-2 world-action model research - 2026-07-13 - Research Publication - The report introduces a latent world-action model that reasons over world dynamics in latent space before generating robot actions. Astribot says Lumo-2 improves over Lumo-1 through predictive latent dynamics, progressive… - source: astribot.com
Astribot publishes Lumo-1 research paper on embodied reasoning and robotic control - 2025-12-09 - Research Publication - Astribot researchers published the Lumo-1 paper describing a generalist vision-language-action model that unifies robot reasoning with robot action. The paper details a three-stage training pipeline spanning embodied… - source: arxiv.org