01
Verifiers
A library of modular components for creating RL environments and training LLM agents.
- Compose environments from datasets, tools and rubrics
- Rubrics synthesized from production failures
- Versioned, shareable environment definitions
Lab · Open source
Environments, trainers and sandboxes. In the open.
The same open components power every PI training loop — inspect them, fork them, or run them yourself.
02
Prime-RL
A framework for asynchronous reinforcement learning (RL) at scale.
- Decoupled trainer, orchestrator and inference
- Scales from one node to multi-node clusters
- Checkpoints deploy straight to your endpoint
03
Sandboxes
For secure code execution optimized for large-scale reinforcement learning.
- Isolated execution for every rollout
- Fast boot for high rollout throughput
- Deterministic, reproducible environments
[ Ready? ]
Your model should get smarter while you sleep.
Engineers, at full stride. Point PI at your checkpoint — first improvement in about 48 hours.