Prime Intellect

@PrimeIntellect

We did it — the first decentralized RL training of a 32B model is complete! Full open-source release is coming in ~1 week, including: checkpoints, data and a detailed technical report.

7:11 PM · Apr 29, 2025