ResearchAreas

Where the hard questions get taken apart and turned into something real.

Reliability first

The receipts

We think the hard part of intelligence is reliability, not capability. So we build systems, then spend most of our time trying to break them.

Our work runs across three areas. How models reason, how they hold up in production, and how we keep them aligned with what people actually want.

Time on evaluation

70

Research areas

03

Models in production

09

Open evaluations

24

How models reach an answer

We study how models actually reach an answer, where that reasoning holds, and where it quietly falls apart.

Models

Systems that hold up at scale

We turn a working model into infrastructure that stays fast and reliable long after the paper ships.

Systems

Alignment you can inspect

We build the evaluations and safeguards that keep a system honest before it ever reaches a real user.

Alignment

Tested against real problems

We take the work out of the benchmark and run it on the messy problems it will actually face in production.

Field tests

What comes out

Proof over promises

What we
prove

Research that stops at a benchmark is easy to fake and hard to trust. Every track here is meant to leave something real behind, whether that is a model, an evaluation, or a system already run in production.