MindForge 教小模型从零写完整软件，不用源码只用可执行文件和文档就把 Qwen3.6-27B 在 ProgramBench 上提升了 11.53%，跟大模型差不多。
论文提出 MindForge 自动流程，将开源命令行程序转为仅暴露编译后可执行文件和文档的源码无关环境。用 GLM-5.2 作为教师智能体生成高质量程序合成轨迹，微调 Qwen3.6-27B。微调后模型在 ProgramBench 上平均测试通过率从 37.98% 提升至 49.51%，并在 7 个未见基准上均获提升，如在 RepoZero-C2Rust 上绝对增益 31.00 分、SWE-bench Verified 上 5.04 分。
MindForge: Teaching Small Language Models Whole-Life-Cycle Software Engineering via Source-Free Program Synthesis
Coding agents have made substantial progress on software engineering tasks that modify existing codebases, including bug fixing and feature implementation. However, constructing a complete program from scratch remains a major challenge: even the frontier models evaluated on ProgramBench fully resolve fewer than 1% of tasks. One obstacle is the lack of scalable training environments for this from-scratch setting, spanning the whole software engineering life cycle, as existing environment-construction frameworks focus only on a single phase in software development. To address this gap, we introduce MindForge, an automated pipeline that converts open-source command-line programs into source-free environments that expose only a compiled reference executable and its documentation. Using MindForge, we construct training environments from repositories disjoint from those in ProgramBench, and curate a high-quality data recipe consisting of program synthesis trajectories using GLM-5.2 as the teacher agent. Fine-tuning Qwen3.6-27B on these trajectories increases its ProgramBench average test pass rate from 37.98% to 49.51%, achieving performance comparable to substantially larger frontier models. Moreover, the fine-tuned model consistently improves over the base model across all seven unseen software engineering benchmarks, spanning long-horizon repository generation and translation, bug fixing, feature implementation, and cross-language issue resolution, with absolute gains of 31.00 points on RepoZero-C2Rust, 14.16 on DeepSWE, 10.70/4.56 on NL2Repo-Bench (with/without tests), 5.04 on SWE-bench Verified, 5.93 on SWE-bench Pro, 5.22 on SWE-bench Multilingual, and 4.94 on FeatBench.