We're a research group pushing the frontier of artificial intelligence — and making it useful, affordable, and open to the world. Not the few. Everyone.
The most capable AI is being concentrated in the hands of a few. We exist to change that — building frontier systems that reach everyone, in every language, on every device, at a price the world can afford.
We do original research at the capability frontier — and publish what we learn instead of locking it away.
A breakthrough only matters when a teacher, a clinic, or a founder halfway across the world can actually use it.
Broad access demands broad responsibility. Safety and alignment are built into the work, not bolted on after.
Three bets on how frontier intelligence reaches the next billion people.
Compressing frontier-grade reasoning into models small and cheap enough to run anywhere — from a data center to a low-cost phone.
Core researchSystems that plan, use tools, and act reliably — so AI becomes a capable collaborator, not just an autocomplete.
Core researchInterpretability, evaluation, and control research that keeps broadly-deployed AI honest, steerable, and safe.
Core researchMultilingual models, open weights and datasets, and free tooling that put the frontier directly into people's hands.
AppliedWhat we're publishing, releasing, and learning in the open.
An intimate technical session on designing systems that automatically prompt coding agents rather than driving them one turn at a time — loop fundamentals, failure modes, and the shift in where the leverage lives.
A small group from research, product, and open source — united by one goal.
Co-Founder
Co-Founder
Co-Founder
We're partnering with investors, institutions, and mission-aligned teams who want frontier AI to reach everyone. Let's talk.