Towards long horizon capabilities
General Reasoning is a new AI R&D company focused on building infrastructure for long horizon AI capabilities.
Our Mission
The next generation of AI models will build entire companies and technologies autonomously over long horizons.
General Reasoning is building the infrastructure to support this future, so agents can be trained and evaluated reliably at scale. We work with industry-leading clients to ensure we can deliver long horizon capabilities safely and reliably.
Latest
For the past six months, we have been hard at work building for the future of autonomous agents.
We began as a team of researchers who had just left frontier labs, and asked a couple of simple questions:
- What does the future of the economy look like in five years time, and how do we build for that today?
- How do we build a lasting company in the ecosystem that outlives the current hype cycle?
- How do we build a future that empowers more people and avoids concentration of power?
We are excited to share with you what we have been working on in the opening quarter of 2026.
Scaling RL Compute