Open Models Can Now Be Trained with RL in Multiple AI Harnesses Simultaneously

Core summary

One sentence to understand this update

Ben Burtenshaw announced that developers can now train open models using Reinforcement Learning (RL) within various AI harnesses like claude code, codex, and opencode, all at once, leveraging openenv's capture proxy.

Impact & opportunity

What this could mean

This advancement provides builders with greater flexibility and efficiency in training and experimenting with RL-driven open models across different environments.