Recently, the frontier of Large Language Model (LLM) capabilities has shifted from single-turn code generation to agentic software engineering-a paradigm where models autonomously navigate, edit, and test complex repositories. While post-training methods have become the de facto approach for code agents, **agentic mid-training**-mid-training (MT) on large-scale data that mirrors authentic agentic workflows-remains critically underexplored due to substantial resource requirements, despite offering a more scalable path to instilling foundational agentic behaviors than relying solely on expensive reinforcement learning. A central challenge in realizing effective agentic mid-training is the distribution mismatch between static training data and the dynamic, feedback-rich environment of real development. To address this, we present a systematic study of agentic mid-training, establishing both the data synthesis principles and training methodology for effective agent development at scale. Central to our approach is **agent-native data**-supervision comprising two complementary types of trajectories: **contextually-native trajectories** that preserve the complete information flow an agent experiences, offering broad coverage and diversity; and **environmentally-native trajectories** collected from executable repositories where observations stem from actual tool invocations and test executions, providing depth and interaction authenticity. We verify the model’s agentic capabilities on ‘SWE-Bench Verified‘. We demonstrate our superiority over the previous open software engineering mid-training recipe ‘Kimi-Dev‘ under two post-training settings with an aligned base model and agentic scaffold, while using less than half mid-training tokens (73.1B). Besides relative advantage, our best performing 32B and 72B models achieve **56.1%** and **58.5%** resolution rates, respectively, which are ...
@misc{zeng_DaVinciDevAgentnativeMidtraining_2026,title={{{daVinci-Dev}}: {{Agent-native Mid-training}} for {{Software Engineering}}},shorttitle={{{daVinci-Dev}}},author={Zeng, Ji and Fu, Dayuan and Mi, Tiantian and Zhuang, Yumin and Huang, Yaxing and Li, Xuefeng and Ye, Lyumanshan and Xie, Muhang and Hua, Qishuo and Huang, Zhen and Jiang, Mohan and Wang, Hanning and Lin, Jifan and Xiao, Yang and Sun, Jie and Wu, Yunze and Liu, Pengfei},year={2026},month=jan,number={arXiv:2601.18418},primaryclass={cs},publisher={arXiv},doi={10.48550/arXiv.2601.18418},archiveprefix={arXiv},langid={american},keywords={Computer Science - Artificial Intelligence,Computer Science - Software Engineering},}
2025
arXiv
LIMI: Less Is More for Agency
Yang
Xiao, Mohan
Jiang, Jie
Sun, and
18 more authors
We define Agency as the emergent capacity of AI systems to function as autonomous agents actively discovering problems, formulating hypotheses, and executing solutions through self-directed engagement with environments and tools. This fundamental capability marks the dawn of the Age of AI Agency, driven by a critical industry shift: the urgent need for AI systems that don’t just think, but work. While current AI excels at reasoning and generating responses, industries demand autonomous agents that can execute tasks, operate tools, and drive real-world outcomes. As agentic intelligence becomes the defining characteristic separating cognitive systems from productive workers, efficiently cultivating machine autonomy becomes paramount. Current approaches assume that more data yields better agency, following traditional scaling laws from language modeling. We fundamentally challenge this paradigm. LIMI (Less Is More for Intelligent Agency) demonstrates that agency follows radically different development principles. Through strategic focus on collaborative software development and scientific research workflows, we show that sophisticated agentic intelligence can emerge from minimal but strategically curated demonstrations of autonomous behavior. Using only 78 carefully designed training samples, LIMI achieves 73.5% on comprehensive agency benchmarks, dramatically outperforming state-of-the-art models: Kimi-K2-Instruct (24.1%), DeepSeek-V3.1 (11.9%), Qwen3-235B-A22B-Instruct (27.5%), and GLM-4.5 (45.1%). Most strikingly, LIMI demonstrates 53.7% improvement over models trained on 10,000 samples-achieving superior agentic intelligence with 128 times fewer samples. Our findings establish the Agency Efficiency Principle: machine autonomy emerges not from data abundance but from strategic curation of high-quality agentic demonstrations.
@misc{LIMILessMore_2025,title={{{LIMI}}: {{Less}} Is {{More}} for {{Agency}}},shorttitle={{{LIMI}}},author={Xiao, Yang and Jiang, Mohan and Sun, Jie and Li, Keyu and Lin, Jifan and Zhuang, Yumin and Zeng, Ji and Xia, Shijie and Hua, Qishuo and Li, Xuefeng and Cai, Xiaojie and Wang, Tongyu and Zhang, Yue and Liu, Liming and Wu, Xia and Hou, Jinlong and Cheng, Yuan and Li, Wenjie and Wang, Xiang and Wang, Dequan and Liu, Pengfei},year={2025},month=sep,number={arXiv:2509.17567},primaryclass={cs},publisher={arXiv},doi={10.48550/arXiv.2509.17567},archiveprefix={arXiv},langid={american},keywords={Computer Science - Artificial Intelligence},}