Introducing OpenHands LM 32B -- A Strong, Open Coding Agent Model
What is OpenHands LM?
OpenHands LM is built on the foundation of Qwen Coder 2.5 Instruct 32B
we use an RL-based framework SWE-Gym and All Hands's scalable training infrastructure a 128K token context window
https://www.all-hands.dev/assets/blog/20250331-openhands-lm-release/performance_scatter.png
パラメタサイズ20倍のDeepSeekと性能は並ぶ