MagicAgent: Towards Generalized Agent Planning
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912934127468544 |
|---|---|
| author | Ren, Xuhui Dong, Shaokang Yang, Chen Gao, Qing Zhao, Yunbin Liu, Yongsheng Geng, Xinwei Li, Xiang Yan, Demei Li, Yanqing Huang, Chenhao Zhu, Dingwei Ye, Junjie Yue, Boxuan Fu, Yingnan Lv, Mengzhe Feng, Zezeng Zhou, Boshen Wang, Bocheng Huang, Xuanjing Jiang, Yu-Gang Gui, Tao Zhang, Qi Zhang, Yunke |
| author_facet | Ren, Xuhui Dong, Shaokang Yang, Chen Gao, Qing Zhao, Yunbin Liu, Yongsheng Geng, Xinwei Li, Xiang Yan, Demei Li, Yanqing Huang, Chenhao Zhu, Dingwei Ye, Junjie Yue, Boxuan Fu, Yingnan Lv, Mengzhe Feng, Zezeng Zhou, Boshen Wang, Bocheng Huang, Xuanjing Jiang, Yu-Gang Gui, Tao Zhang, Qi Zhang, Yunke |
| contents | The evolution of Large Language Models (LLMs) from passive text processors to autonomous agents has established planning as a core component of modern intelligence. However, achieving generalized planning remains elusive, not only by the scarcity of high-quality interaction data but also by inherent conflicts across heterogeneous planning tasks. These challenges result in models that excel at isolated tasks yet struggle to generalize, while existing multi-task training attempts suffer from gradient interference. In this paper, we present \textbf{MagicAgent}, a series of foundation models specifically designed for generalized agent planning. We introduce a lightweight and scalable synthetic data framework that generates high-quality trajectories across diverse planning tasks, including hierarchical task decomposition, tool-augmented planning, multi-constraint scheduling, procedural logic orchestration, and long-horizon tool execution. To mitigate training conflicts, we propose a two-stage training paradigm comprising supervised fine-tuning followed by multi-objective reinforcement learning over both static datasets and dynamic environments. Empirical results show that MagicAgent-32B and MagicAgent-30B-A3B achieve superior performance across diverse open-source benchmarks (\emph{e.g.}, $75.1\%$ on Worfbench and $86.9\%$ on BFCL-v3), as well as strong results on our in-house MagicEval benchmarks, substantially outperforming existing sub-100B models and surpassing leading ultra-scale models, including GPT-5.2, Kimi-K2 and GLM-4.7. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_19000 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MagicAgent: Towards Generalized Agent Planning Ren, Xuhui Dong, Shaokang Yang, Chen Gao, Qing Zhao, Yunbin Liu, Yongsheng Geng, Xinwei Li, Xiang Yan, Demei Li, Yanqing Huang, Chenhao Zhu, Dingwei Ye, Junjie Yue, Boxuan Fu, Yingnan Lv, Mengzhe Feng, Zezeng Zhou, Boshen Wang, Bocheng Huang, Xuanjing Jiang, Yu-Gang Gui, Tao Zhang, Qi Zhang, Yunke Artificial Intelligence Human-Computer Interaction The evolution of Large Language Models (LLMs) from passive text processors to autonomous agents has established planning as a core component of modern intelligence. However, achieving generalized planning remains elusive, not only by the scarcity of high-quality interaction data but also by inherent conflicts across heterogeneous planning tasks. These challenges result in models that excel at isolated tasks yet struggle to generalize, while existing multi-task training attempts suffer from gradient interference. In this paper, we present \textbf{MagicAgent}, a series of foundation models specifically designed for generalized agent planning. We introduce a lightweight and scalable synthetic data framework that generates high-quality trajectories across diverse planning tasks, including hierarchical task decomposition, tool-augmented planning, multi-constraint scheduling, procedural logic orchestration, and long-horizon tool execution. To mitigate training conflicts, we propose a two-stage training paradigm comprising supervised fine-tuning followed by multi-objective reinforcement learning over both static datasets and dynamic environments. Empirical results show that MagicAgent-32B and MagicAgent-30B-A3B achieve superior performance across diverse open-source benchmarks (\emph{e.g.}, $75.1\%$ on Worfbench and $86.9\%$ on BFCL-v3), as well as strong results on our in-house MagicEval benchmarks, substantially outperforming existing sub-100B models and surpassing leading ultra-scale models, including GPT-5.2, Kimi-K2 and GLM-4.7. |
| title | MagicAgent: Towards Generalized Agent Planning |
| topic | Artificial Intelligence Human-Computer Interaction |
| url | https://arxiv.org/abs/2602.19000 |