35B大模型通过自进化框架实现机器学习工程的持续优化，性能超越GPT-5.5。
Frontis-MA1: Training an AI4AI Model towards Recursive Self-Improvement in Machine Learning Engineering
- 用四类程序演化操作构建可执行的自进化系统
- 12小时任务预算下准确率从39.39%提升至71.21%
- 开源全栈工具链支持可复现的AI4AI研究
递归自我改进（RSI）需要能优化自身构建过程的AI系统（即AI4AI），机器学习工程（MLE）为此提供了可执行的测试平台。我们提出OpenMLE，一个开放的全栈系统，涵盖可验证的任务环境（OpenMLE-Gym）、操作学习（OpenMLE-RL）与长周期搜索（OpenMLE-Evo）。在此基础上，我们将Frontis-MA1（35B）作为元演化代理进行后训练，围绕四个原子程序演化操作（起草、改进、调试、交叉）进行对齐。这些操作通过执行驱动的SFT和RL在去重数据上训练，再组合成长周期搜索，实现学习与进化的闭环。在12小时任务预算、单张RTX 4090（12GB VRAM）限制下，前端结果在MLE-Bench Lite上将奖牌平均分从39.39%提升至60.61%（使用OpenMLE-Evo），达到71.21%（使用OpenMLE-Evo-Max），超过GPT-5.5+Codex，接近GPT-5.6 Sol和2.8T Kimi K3。在未见的NatureBench Lite上，组件可迁移：固定框架换模型，匹配最优解从50%升至70%；固定模型换搜索器，从20%升至50%。我们公开模型权重与完整OpenMLE栈，推动可复现的可执行AI4AI研究。代码：https://github.com/FrontisAI/OpenRSI
原文摘要 · Abstract (English)
Recursive self-improvement (RSI) requires AI systems that improve the process of building AI (i.e., AI4AI); machine learning engineering (MLE) offers a concrete, executable testbed for studying this capability. We introduce OpenMLE, an open full-stack system for RSI research in MLE, spanning verifiable task environments with execution feedback (OpenMLE-Gym), operator learning (OpenMLE-RL), and long-horizon search (OpenMLE-Evo). On this stack we post-train Frontis-MA1 (35B) as a meta-evolution agent for MLE, aligning post-training and inference around four atomic program-evolution operators (Draft, Improve, Debug, Crossover): the same operators are trained via execution-grounded SFT and RL on data deduplicated against all evaluation benchmarks, then composed into long-horizon search, coupling learning and evolution in a single loop. On MLE-Bench Lite under a 12-hour per-task budget on one RTX 4090 capped at 12 GB VRAM, Frontis-MA1 (35B) improves Medal Average from 39.39% to 60.61% over its base model with OpenMLE-Evo, and reaches 71.21% with OpenMLE-Evo-Max (benchmark-independent experience priors and asynchronous search), exceeding GPT-5.5 + Codex and approaching GPT-5.6 Sol and the 2.8T Kimi K3. On held-out NatureBench Lite, both components transfer: with the framework fixed, swapping in the trained model raises Match-SOTA from 50% to 70%; with the model fixed, swapping in OpenMLE-Evo raises it from 20% to 50%. We release the model weights and the full OpenMLE stack to enable reproducible research on executable AI4AI toward RSI. Code: https://github.com/FrontisAI/OpenRSI
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。