MindForge: Teaching Small Language Models Whole-Life-Cycle Software Engineering via Source-Free Program Synthesis

• URL: http://arxiv.org/abs/2607.27146v1
• Date: Wed, 29 Jul 2026 17:23:02 GMT
• Title: MindForge: Teaching Small Language Models Whole-Life-Cycle Software Engineering via Source-Free Program Synthesis
• Abstract summary: We introduce MindForge, an automated pipeline that converts open-source command-line programs into source-free environments.<n>We construct training environments from repositories disjoint from those in ProgramBench, and curate a high-quality data recipe.<n>Fine-tuning Qwen3.6-27B on these trajectories increases its ProgramBench average test pass rate from 37.98% to 49.51%.
• Score: 17.260324148993046
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: 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.

Related papers

• UI-Mate: Advancing Open-Weight Foundation GUI Agents with In-Context Demonstrations [68.29650478987561]
  UI-Mate is a foundation GUI agent that integrates an environment-grounded training stack with in-context demonstration learning.<n>It sets a new open-weight state of the art on general computer-use benchmarks, scoring 77.0% on OSWorld and 66.2% on WindowsAgentArena.
  arXiv  Detail & Related papers  (2026-08-16T20:59:59Z)
• LoopsBench: From Harness Engineering to Loop Engineering in Benchmarking Coding Agent [60.27239351179265]
  LOOPSBENCH is a benchmark for loop engineering in coding agent evaluation.<n>It comprises 112 tasks from authentic sources spanning 8 programming languages and 9 domains.<n>Its flow-aware runtime releases tests along the ready frontier and retains completed nodes as regression obligations.
  arXiv  Detail & Related papers  (2026-07-31T20:18:25Z)
• KAT-Coder-V2.5 Technical Report [61.907486544595336]
  KAT-Coder-V2.5 is a coding-focused agentic model trained to act autonomously inside real, executable repositories.<n>It reconstructs multilingual repositories into sandboxed environments with fail-to-pass and pass-to-pass verification at scale.<n>We further scale reinforcement learning with harness randomization, a reliability-hardened sandbox, an asymmetric actor--critic PPO with hindsight-augmented value estimation, and a harness-oriented reward framework.
  arXiv  Detail & Related papers  (2026-07-06T08:14:02Z)
• ProCUA-SFT Technical Report [80.97543110323542]
  Training computer-use agents (CUAs) requires large-scale, diverse trajectory data collected in full desktop environments.<n>We present ProCUA-SFT, a dataset of 3.1M step-level SFT samples distilled from 93K synthetic trajectories.<n>Fine-tuning UI-TARS 7B on ProCUA-SFT for one epoch yields 45.0% on OSWorld -- an 18.7 percentage-point improvement over the base model.
  arXiv  Detail & Related papers  (2026-06-15T22:04:11Z)
• Open-SWE-Traces: Advancing Dual-Mode Multilingual Distillation for Software Engineering Agents [42.7127695634038]
  ourdataset is a dataset of 207,489 agentic trajectories spanning nine programming languages.<n>This data facilitates the training of models capable of long-horizon reasoning.<n>The best performing model achieves resolve rates of 61.7% on SWE-bench Verified, 57.1% on SWE-bench Multilingual, and 36.8% on SWE-bench Pro.
  arXiv  Detail & Related papers  (2026-06-14T22:10:06Z)
• Fine-Tuning GPT-5 for GPU Kernel Generation [5.109141377873154]
  We present Makora's environment and tools for reinforcement learning finetuning of frontier models.<n>In the single-attempt setting, our fine-tuned model improves kernel correctness from 43.7% to 77.0%.<n>When integrated into a full coding agent, it is able to solve up to 97.4% of problems in an expanded KernelBench suite.
  arXiv  Detail & Related papers  (2026-02-11T16:22:54Z)
• SWE-World: Building Software Engineering Agents in Docker-Free Environments [91.17484806743641]
  SWE-World is a Docker-free framework that replaces physical execution environments with a learned surrogate for training and evaluating software engineering agents.<n>We show that SWE-World raises Qwen2.5-Coder-32B from 6.2% to 52.0% via Docker-free SFT, 55.0% with Docker-free RL, and 68.2% with further TTS.
  arXiv  Detail & Related papers  (2026-02-03T11:44:39Z)
• SWE-Bench++: A Framework for the Scalable Generation of Software Engineering Benchmarks from Open-Source Repositories [2.951332247539421]
  We introduce SWE-Bench++, an automated framework that generates repository-level coding tasks from open-source GitHub projects.<n>Unlike synthetic approaches, our pipeline harvests live pull requests to cover both bug fixes and feature requests across 11 languages.<n>Our initial benchmark consists of 11,133 instances from 3,971 repositories across 11 languages.
  arXiv  Detail & Related papers  (2025-12-19T10:16:51Z)
• Lifecycle-Aware code generation: Leveraging Software Engineering Phases in LLMs [12.70863561286374]
  We introduce a lifecycle-aware framework that incorporates intermediate artifacts into both the training and inference stages.<n> Experiments show that lifecycle-level fine-tuning improves code correctness by up to 75% over the same model before fine-tuning.<n>Open-source LLMs, once fine-tuned under our framework, match or slightly outperform models pretrained on code.
  arXiv  Detail & Related papers  (2025-10-28T02:54:02Z)
• Web-Bench: A LLM Code Benchmark Based on Web Standards and Frameworks [6.303398878886812]
  We propose a new benchmark, Web-Bench, which contains 50 projects, each consisting of 20 tasks with sequential dependencies.<n>We aim to cover the foundational elements of Web development: Web Standards and Web Frameworks.<n>On our given benchmark agent (Web-Agent), SOTA (Claude 3.7 Sonnet) achieves only 25.1% Pass@1, significantly lower (better) than SWE-Bench's Verified (65.4%) and Full (33.8%) scores.
  arXiv  Detail & Related papers  (2025-05-12T12:06:23Z)
• Lingma SWE-GPT: An Open Development-Process-Centric Language Model for Automated Software Improvement [62.94719119451089]
  Lingma SWE-GPT series learns from and simulating real-world code submission activities. Lingma SWE-GPT 72B resolves 30.20% of GitHub issues, marking a significant improvement in automatic issue resolution.
  arXiv  Detail & Related papers  (2024-11-01T14:27:16Z)
• Building Math Agents with Multi-Turn Iterative Preference Learning [56.71330214021884]
  This paper studies the complementary direct preference learning approach to further improve model performance.<n>Existing direct preference learning algorithms are originally designed for the single-turn chat task.<n>We introduce a multi-turn direct preference learning framework, tailored for this context.
  arXiv  Detail & Related papers  (2024-09-04T02:41:04Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.