Generalist AI Inc. unveiled its GEN-1 general-purpose AI model for robotics, achieving 99% success rates on tasks where previous models reached 64%. The model completes tasks three times faster and requires only one hour of robot data per task. Founded in 2024 in San Mateo, California, Generalist builds embodied foundation models for versatile robots. GEN-1 builds on the 2024 GEN-0 model, scaling with half a million hours of real-world data and algorithmic advances. Unlike prior models relying on costly teleoperation datasets, GEN-1 trains primarily on human activity data from low-cost wearables. The model demonstrates rapid learning, adaptability to new environments, and physical common sense, improving reliability and precision. It achieves comparable performance to GEN-0 with ten times less task-specific data and fine-tuning. Generalist claims GEN-1 unlocks commercial viability across diverse applications but notes some tasks require success rates above 99% for real-world use. The company will highlight embodied and physical AI developments at the 2026 Robotics Summit & Expo in Boston on May 27-28. AECM professionals should monitor GEN-1’s integration potential for robotics automation, which promises faster, more reliable task execution with reduced data needs, impacting manufacturing and construction automation workflows.
Source: source. Read the original story →
