Youtu-LLM: Unlocking the Native Agentic Potential for Lightweight Large Language Models

• URL: http://arxiv.org/abs/2512.24618v2
• Date: Mon, 05 Jan 2026 02:44:28 GMT
• Title: Youtu-LLM: Unlocking the Native Agentic Potential for Lightweight Large Language Models
• Authors: Junru Lu, Jiarui Qin, Lingfeng Qiao, Yinghui Li, Xinyi Dai, Bo Ke, Jianfeng He, Ruizhi Qiao, Di Yin, Xing Sun, Yunsheng Wu, Yinsong Liu, Shuangyin Liu, Mingkong Tang, Haodong Lin, Jiayi Kuang, Fanxu Meng, Xiaojuan Tang, Yunjia Xi, Junjie Huang, Haotong Yang, Zhenyi Shen, Yangning Li, Qianwen Zhang, Yifei Yu, Siyu An, Junnan Dong, Qiufeng Wang, Jie Wang, Keyu Chen, Wei Wen, Taian Guo, Zhifeng Shen, Daohai Yu, Jiahao Li, Ke Li, Zongyi Li, Xiaoyu Tan,
• Abstract summary: We introduce Youtu-LLM, a lightweight language model that harmonizes high computational efficiency with native agentic intelligence.<n>Youtu-LLM is pre-trained from scratch to systematically cultivate reasoning and planning capabilities.
• Score: 78.73992315826035
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We introduce Youtu-LLM, a lightweight yet powerful language model that harmonizes high computational efficiency with native agentic intelligence. Unlike typical small models that rely on distillation, Youtu-LLM (1.96B) is pre-trained from scratch to systematically cultivate reasoning and planning capabilities. The key technical advancements are as follows: (1) Compact Architecture with Long-Context Support: Built on a dense Multi-Latent Attention (MLA) architecture with a novel STEM-oriented vocabulary, Youtu-LLM supports a 128k context window. This design enables robust long-context reasoning and state tracking within a minimal memory footprint, making it ideal for long-horizon agent and reasoning tasks. (2) Principled "Commonsense-STEM-Agent" Curriculum: We curated a massive corpus of approximately 11T tokens and implemented a multi-stage training strategy. By progressively shifting the pre-training data distribution from general commonsense to complex STEM and agentic tasks, we ensure the model acquires deep cognitive abilities rather than superficial alignment. (3) Scalable Agentic Mid-training: Specifically for the agentic mid-training, we employ diverse data construction schemes to synthesize rich and varied trajectories across math, coding, and tool-use domains. This high-quality data enables the model to internalize planning and reflection behaviors effectively. Extensive evaluations show that Youtu-LLM sets a new state-of-the-art for sub-2B LLMs. On general benchmarks, it achieves competitive performance against larger models, while on agent-specific tasks, it significantly surpasses existing SOTA baselines, demonstrating that lightweight models can possess strong intrinsic agentic capabilities.

Related papers

• MagicAgent: Towards Generalized Agent Planning [73.21129030631421]
  We present textbfMagicAgent, a series of foundation models specifically designed for generalized agent planning.<n>We introduce a lightweight and scalable synthetic data framework that generates high-quality trajectories across diverse planning tasks.<n>We show that MagicAgent-32B and MagicAgent-30B-A3B achieve superior performance across diverse open-source benchmarks.
  arXiv  Detail & Related papers  (2026-02-22T01:39:16Z)
• SimuAgent: An LLM-Based Simulink Modeling Assistant Enhanced with Reinforcement Learning [3.1436750864792375]
  We introduce SimuAgent, an LLM-powered modeling and simulation agent tailored for Simulink.<n>SimuAgent replaces XML with a concise, dictionary-style Python representation, dramatically cutting token counts.<n>A lightweight plan-execute architecture, trained in two stages, equips the agent with both low-level tool skills and high-level design reasoning.
  arXiv  Detail & Related papers  (2026-01-08T18:10:35Z)
• PRInTS: Reward Modeling for Long-Horizon Information Seeking [74.14496236655911]
  We introduce PRInTS, a generative PRM trained with dual capabilities.<n>We show that PRInTS enhances information-seeking abilities of open-source models as well as specialized agents.
  arXiv  Detail & Related papers  (2025-11-24T17:09:43Z)
• Agent-Omni: Test-Time Multimodal Reasoning via Model Coordination for Understanding Anything [12.274140974616747]
  Multimodal large language models (MLLMs) have shown strong capabilities but remain limited to fixed modality pairs.<n>We propose an Agent- Omni framework that coordinates existing foundation models through a master-agent system.
  arXiv  Detail & Related papers  (2025-11-04T18:59:09Z)
• Generative World Models of Tasks: LLM-Driven Hierarchical Scaffolding for Embodied Agents [0.0]
  We propose an effective world model for decision-making that models the world's physics and its task semantics.<n>A systematic review of 2024 research in low-resource multi-agent soccer reveals a clear trend towards integrating symbolic and hierarchical methods.<n>We formalize this trend into a framework for Hierarchical Task Environments (HTEs), which are essential for bridging the gap between simple, reactive behaviors and sophisticated, strategic team play.
  arXiv  Detail & Related papers  (2025-09-05T01:03:51Z)
• Foundation Model for Skeleton-Based Human Action Understanding [56.89025287217221]
  This paper presents a Unified Skeleton-based Dense Representation Learning framework.<n>USDRL consists of a Transformer-based Dense Spatio-Temporal (DSTE), Multi-Grained Feature Decorrelation (MG-FD), and Multi-Perspective Consistency Training (MPCT)
  arXiv  Detail & Related papers  (2025-08-18T02:42:16Z)
• Speed Always Wins: A Survey on Efficient Architectures for Large Language Models [51.817121227562964]
  Large Language Models (LLMs) have delivered impressive results in language understanding, generation, reasoning, and pushes the ability boundary of multimodal models.<n> Transformer models, as the foundation of modern LLMs, offer a strong baseline with excellent scaling properties.<n>The traditional transformer architecture requires substantial computations and poses significant obstacles for large-scale training and practical deployment.
  arXiv  Detail & Related papers  (2025-08-13T14:13:46Z)
• OptMerge: Unifying Multimodal LLM Capabilities and Modalities via Model Merging [124.91183814854126]
  Model merging seeks to combine multiple expert models into a single model.<n>We introduce a benchmark for model merging research that clearly divides the tasks for MLLM training and evaluation.<n>We find that model merging offers a promising way for building improved MLLMs without requiring training data.
  arXiv  Detail & Related papers  (2025-05-26T12:23:14Z)
• Scaling Laws for Native Multimodal Models [53.490942903659565]
  We revisit the architectural design of native multimodal models and conduct an extensive scaling laws study.<n>Our investigation reveals no inherent advantage to late-fusion architectures over early-fusion ones.<n>We show that incorporating Mixture of Experts (MoEs) allows models to learn modality-specific weights, significantly benefiting performance.
  arXiv  Detail & Related papers  (2025-04-10T17:57:28Z)
• Factored Agents: Decoupling In-Context Learning and Memorization for Robust Tool Use [4.437184840125514]
  We propose a novel factored agent architecture designed to overcome the limitations of traditional single-agent systems in agentic AI.<n>Our approach decomposes the agent into two specialized components: (1) a large language model that serves as a high level planner and in-context learner, and (2) a smaller language model which acts as a memorizer of tool format and output.<n> Empirical evaluations demonstrate that our factored architecture significantly improves planning accuracy and error resilience, while elucidating the inherent trade-off between in-context learning and static memorization.
  arXiv  Detail & Related papers  (2025-03-29T01:27:11Z)
• MindLLM: Pre-training Lightweight Large Language Model from Scratch, Evaluations and Domain Applications [46.337078949637345]
  We present MindLLM, a novel series of bilingual lightweight large language models, trained from scratch. A thorough account of experiences accrued during large model development is given, covering every step of the process. MindLLM consistently matches or surpasses the performance of other open-source larger models on some public benchmarks.
  arXiv  Detail & Related papers  (2023-10-24T12:22:34Z)

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.