We build intelligence that learns from experience and acts in the world. Our research connects perception, decision-making, memory and action.
Reflex-1 is our 421M-parameter decision model, with public weights and CPU/GPU inference. Samsara-VLA is our experimental work on compact robot policies with memory. Colony brings memory and tools to persistent agents. re1 keeps recorded observations and captured places available for physical AI.
Explore models, demonstrations and research on goodailabs.com.
