LMSYS, AMD and Miles train Qwen3-30B-A3B SWE agent on MI355X

AILMSYS Org, with AMD and the Miles team, trained Qwen3-30B-A3B as a software engineering agent on AMD MI355X using fully asynchronous agentic RL. After SFT and RL, SWE-bench Verified pass@1 rose from 7.3% to 27.5% on a 100-task subset with four attempts per task. RL steps dropped from 850s to 284s, and rollout throughput rose about 4x, to roughly 3,480 trajectories per hour with 384 sandboxes running at once.

LMSYS OrgOfficial

@lmsysorg

View on X

🚀 Training SWE Agents with Miles on AMD MI355X

The @AIatAMD and Miles teams trained Qwen3-30B-A3B with fully asynchronous agentic RL using SGLang, Megatron-LM, and Harbor. Keeping long trajectories and hundreds of CPU sandboxes from stalling training took some engineering.

Highlights:
- SWE-bench Verified pass@1: 7.3% → 27.5% after SFT + RL (100-task stratified subset, 4 attempts per task)
- 3× faster RL steps: 850s → 284s
- ~4× rollout throughput (~860 → ~3,480 trajectories/h), with 384 agent sandboxes running at once.

Full report: http://lmsys.org/blog/2026-10-22-miles-mi355x-agentic-rl

Source details

LMSYS Org · x.comPublished Oct 28, 5:01 PM ET · added here 5:09 PM ET

Original title: 🚀 Training SWE Agents with Miles on AMD MI355X

LMSYS, AMD and Miles train Qwen3-30B-A3B SWE agent on MI355X

Topics