IFM Open-Sources RL360, Massive Agentic RL Infrastructure Behind K2 Horizon

IFM has open-sourced RL360, the RL infrastructure used to train the K2 Horizon models (375B total, 23B active), supporting long-horizon agentic RL with up to 8192 concurrent sandboxes and around 1.3 million training runs.

2026-10-23 ~ 2026-10-23 · 2 related posts