发生了什么
Clément Delangue 介绍了 Qwen3.8-27B-pi-adapted,这是一个通过强化学习和 LoRA 适配器为 pi 编程智能体适配的 Qwen 模型。文章报告称,在 SWE-bench Pro (Hard 51) 上,pass@1 从 29.4% 提升至 62.7%,在 Terminal-Bench 2.1 上从 39.5% 提升至 55.8%,且每个任务的工具调用次数减少了 55–62%。开放权重和方法论已提供链接。
推荐理由
该适配模型提升了在 SWE-bench Pro (Hard 51) 和 Terminal-Bench 2.1 上的 pass@1,同时每个任务的工具调用次数减少了 55–62%,这表明编程智能体可以从针对其特定工具进行调优的模型中受益。
Introducing Qwen3.8-27B-pi-adapted 🤗
We adapted Qwen's model specifically for the pi coding agent using reinforcement learning and a LoRA adapter.
The results, measured as pass@1:
🚀 SWE-bench Pro (Hard 51): 29.4% → 62.7%
💻 Terminal-Bench 2.1: 39.5% → 55.8%
⚡ 55–62% fewer tool calls per task
This is the power of open models: you can adapt them to YOUR tools and YOUR use cases. Let’s build! 💛
Open weights: huggingface.co/harness-hub/Qw…
Methodology: huggingface.co/blog/harness-h…
这件事到目前为止
Harness Hub has released Qwen3.8-27B-pi-adapted, an iteration of the Qwen model optimized for the pi coding agent using reinforcement learning and a LoRA adapter. The model's weights and methodology are available publicly. Clément Delangue reported that this adaptation improves performance across coding benchmarks. On SWE-bench Pro (Hard 51), pass@1 rates increased from 29.4% to 62.7%, while scores on Terminal-Bench 2.1 rose from 39.5% to 55.8%. Additionally, the model reportedly requires 55–62% fewer tool calls per task.
