BGI Labs
AGI first,
superintelligence
next.
BGI Labs is building the next-generation AGI stack, with the cognitive layer that fuels the agentic economy.
Scroll to explore ↓
Defining AGI
“AGI is AI that can learn and succeed at goals it was never trained for, in situations it has never seen.”
Dr. Ben GoertzelExecutive Chairman & Chief Scientist
Our approach
A better path than bigger transformer models.
Large language models are impressive, but the industry is betting almost everything on making them bigger. We’re taking a different path: many kinds of thinking and memory working together in one mind, so the system can keep learning from experience, show how it reaches its answers, and improve over time.
Hyperon + Omega
What we’re building.
Hyperon is the mind behind everything we build: one system where memory, reasoning, and language work together, the way they do in a person. Today’s large language models plug into it as parts, so every advance from the big labs makes it stronger.
Omega is how businesses put that mind to work. It gives them AI agents that remember what they learn, reason through problems, and act within the rules they set, and those agents can work together as a team.
Because Hyperon is neural-symbolic, it also brings new ways to train neural networks. With predictive and causal coding, its networks learn the way the brain does: by predicting what comes next and correcting themselves as they go. That opens the door to continual, transfer, and lifelong learning, which standard neural networks struggle with.
Explore the technologySince OpenCog
20 years of building.
Now it runs at scale.
For 20 years, our research community built and optimized the tools to run this design at scale. BGI Labs was formed to turn that research into products, and Omega is the first.
-
OpenCog
A design for AGI
-
Hyperon
The tools to run it at scale
-
Omega
The series of products
Leadership
The team that
launched the field.
Executive Chairman & Chief Scientist
Dr. Ben Goertzel
Dr. Goertzel launched the term and concept of artificial general intelligence and has led AGI research for more than two decades. He architected OpenCog and Hyperon, has published 260+ papers and 25+ books, and helped create the humanoid robot Sophia at Hanson Robotics.
Chief Executive Officer
Dr. Ben Heng
Dr. Heng brings the playbook for taking breakthrough technology from prototype to pilot to production. He holds 85+ patents, has held senior leadership roles at Tesla, Intel, SolarCity, and ESS, and is General Partner of 99VC, an agentic and physical AI fund.
Research bench
Research leaders shaping the path to AGI
Behind them is a research bench of 20+ senior engineers, scientists, and PhDs working across Hyperon’s reasoning, language, and memory layers. Together, they bring 750+ publications, 9,000+ citations, and two decades of collective work on AGI.
Our commitment
Beneficial,
by design.
General intelligence will reshape the world, so we’re building it to be open to everyone, safe to trust, and made for the benefit of all. We call this beneficial general intelligence. Ethics and safety shape how we design, build, and release everything we make.
