Generalist AI Launches GEN-1.5: A Breakthrough in One-Shot Robot Learning
- Published
- Source
- HumanoidsDaily
Generalist AI has introduced GEN-1.5, an advanced embodied foundation model capable of one-shot learning for dexterous tasks. This model can learn new behaviors with just 3 to 12 seconds of sensorimotor demonstration data, eliminating the need for gradient updates. The model's performance shows a 59% success rate in ten manipulation tasks through one-shot prompting, which can be improved to 83% with fine-tuning.
The significance of GEN-1.5 lies in its ability to adapt quickly to new tasks with minimal training, representing a shift towards real-time adaptability in robotics. The model has been continuously trained for over eight months, leading to improvements in data absorption and compute efficiency. This efficiency is crucial as it reduces the traditionally high computational demands for training robotic models.
Looking ahead, GEN-1.5's compositional generalization allows it to chain distinct tasks into longer skills autonomously. Additionally, its zero-shot sim-to-real transfer capability enables it to perform tasks based solely on simulated prompts. While the company acknowledges that skills learned in-context are less robust than those from fine-tuning, the model's improvisational abilities demonstrate significant advancements in robotic flexibility.
Original report
Generalist AI Unveils GEN-1.5: One-Shot Robot Learning and the End of Heavy Fine-Tuning
HumanoidsDaily · [email protected] (Humanoids Daily Staff)This briefing is an independently written summary based on publicly available reporting and is provided for industry information and news discovery. The original report and source publication are credited and linked where applicable. RobotToday does not claim ownership of third-party source material.
Rights concerns? Contact [email protected] with the relevant URL and details. We will review the matter and take appropriate action where warranted.
More related news
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
-
11
ROBOTTODAY Weekly October 19 – 23, 2026 RobotToday · 1 week ago
- 12
Related Suppliers
Generalist AI
United States
Frontier AI research company building embodied foundation models for general-purpose robots, with GEN-1 achieving 99% task success rates.
Rokid Inc
Hangzhou, China AI and AR glasses maker founded 2014; sells the Rokid Glasses, AI Glasses Style, Max 2 and AR Spatial lines to more than 1 million users.
Kemaro AG
Switzerland
Swiss maker of autonomous, water- and chemical-free dry sweeping robots that clean large industrial floors 24/7.
The Robot Learning Company
United States
YC X25 startup building open-source AI-native robot arms; TRLC-DK1 bimanual dev kit enables imitation learning and teleoperation for around USD 1,200.
Assured Robot Intelligence
United States
San Diego robot-AI start-up founded by Lerrel Pinto and Xiaolong Wang, building humanoid foundation models; acquired by Meta in May 2026.
Acumino
United States
Physical AI startup founded 2021; hardware-agnostic dexterous robot models for industry. Raised $11.7M seed Jun 2026 via Google DeepMind accelerator.
Seed Robotics Team, ByteDance Seed 字节跳动Seed机器人团队
China
ByteDance Seed's robotics research team, developer of the GR-3 vision-language-action model and the ByteMini bimanual mobile robot.
Real Time Automation, Inc.
United States
US-based manufacturer of industrial protocol gateways, edge data historians, and embedded connectivity software for factory floor automation.
Real-Time Innovations (RTI)
United States
RTI is the world's largest DDS middleware vendor, providing RTI Connext software as the backbone for autonomous vehicles, defense, and aerospace systems.
Crusoe
United States
Denver-based AI infrastructure company operating energy-first GPU cloud data centers, including a nuclear-powered facility built with Aalo Atomics.
KAIST RaiLab
Korea, Republic of (South)
KAIST mechanical engineering lab of Prof. Jemin Hwangbo that builds the RAIBO quadrupeds; RAIBO 2 ran a full 42.195 km marathon on one charge.
Lili-o Inc.
Germany
Early-stage robot software startup building a hardware-agnostic deterministic execution engine that teaches humanoids skills from one human demonstration.
Robotics needs a service framework.
RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.
Join the RobotToday community on LinkedIn
Daily robotics news, in-depth analysis, conference highlights, and discussions with professionals worldwide.
