MongoDB - Build AI That Scales
Mini toolsGEN 1
GEN-1 is Generalist AI’s latest embodied foundation model for robotics, built to emit actions in real time rather than just understand scenes or language. The company describes it as a large multimodal system trained from scratch on a dataset of over half a million hours of high-fidelity physical interaction data, with advances in pretraining, reinforcement learning, multimodal human guidance, and inference-time techniques. Generalist says GEN-1 shows 99% average success on selected simple physical tasks, around 3x faster task completion than prior state of the art, and strong data efficiency with about 1 hour of robot data per result.
New
Multimodal
Gen 3
Overview
GEN-1 is Generalist AI’s large multimodal robot foundation model for real-time action generation in the physical world. It is positioned as a general-purpose embodied AI system that reaches “mastery” on simple physical tasks, with reported 99% average success on selected tasks, about 3x faster task completion than prior state of the art, and adaptation using only about 1 hour of robot data.
Tools using GEN 1
No tools found for this model yet.
Last updated: April 3, 2026
