Very interesting new work from Microsoft
This work is related to this emerging theme of leveraging harnesses for model post-training.
Agent Lightning v1.0 connects any harness to RL through an endpoint proxy in about 3,500 lines, then works through what breaks in that setup, retokenization, sample merging, advantage calculation, loss normalization, and backend scheduling.
Using 6K training examples and modest compute, it moves Qwen3.5-9B on SWE-bench Verified from 41.8% to 56.4%.
This work is related to this emerging theme of leveraging harnesses for model post-training.
Agent Lightning v1.0 connects any harness to RL through an endpoint proxy in about 3,500 lines, then works through what breaks in that setup, retokenization, sample merging, advantage calculation, loss normalization, and backend scheduling.
Using 6K training examples and modest compute, it moves Qwen3.5-9B on SWE-bench Verified from 41.8% to 56.4%.