Hugging Face 发布 LoRA 适配器,Qwen3.8 编码成绩从 39% 升至 65%
New LoRA Adapter Lifts Qwen3.8 Coding Success From 39% to 65%
写代码的朋友可以留意,Hugging Face 给 Qwen3.8-27B 出了个 LoRA 插件,专门强化 pi 编码智能体,SWE-bench Pro 成绩从 39% 直接拉到 65%。Apache 2.0 开源,挂上就能用,不用重新训练模型。
Hugging Face 发布了 Qwen3.8-27B-pi-adapted,这是一个 Apache-2.0 许可的 LoRA 适配器,专为 pi 编码智能体训练,采用强化学习方法。在 SWE-bench Pro hard-51 基准上,pass@1 成绩从 39.2% 提升到 64.7%;Terminal-Bench 2.1 成绩从 39.5% 提升到 55.8%。
技术上是给 Qwen3.8-27B 挂载一个小型适配器,而非重新训练整个模型,配合 pi 编码智能体使用。由于采用 Apache-2.0 许可,开发者可以直接下载挂载,改动成本很低。
对使用 Qwen3.8-27B 做编程任务的开发者来说,接近翻倍的基准提升意味着更低成本获得更强的代码修复能力,也可以和其他开源编程模型方案做对比参考。
New LoRA Adapter Lifts Qwen3.8 Coding Success From 39% to 65%
Hugging Face released Qwen3.8-27B-pi-adapted, an Apache-2.0 LoRA adapter trained with reinforcement learning for the pi coding agent. It lifts SWE-bench Pro hard-51 pass@1 from 39.2% to 64.7% and Terminal-Bench 2.1 from 39.5% to 55.8%. The post New LoRA Adapter Lifts Qwen3.8 Coding Success From 39% to 65% appeared first on MarkTechPost .
This report is categorized under AITOP's "Models" desk, focusing on China's AI industry and domestic foundation model ecosystem. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Qwen3.8-27B, Hugging Face, LoRA, SWE-bench.
Primary source verification: published directly by marktechpost with first-party authority.