Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2511.11793 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911612385886208 |
|---|---|
| author | MiroMind Team Bai, Song Bing, Lidong Chen, Carson Chen, Guanzheng Chen, Yuntao Chen, Zhe Chen, Ziyi Dai, Jifeng Dong, Xuan Dou, Wenhan Deng, Yue Fu, Yunjie Ge, Junqi Han, Chenxia Huang, Tammy Huang, Zhenhang Jiao, Jerry Jiang, Shilei Jiao, Tianyu Jian, Xiaoqi Lei, Lei Li, Ruilin Luo, Gen Li, Tiantong Lin, Xiang Liu, Ziyuan Li, Zhiqi Ni, Jie Ren, Qiang Sun, Pax Su, Shiqian Tao, Chenxin Wang, Bin Wang, Wenhai Wang, Haonan Wang, James Wang, Jin Wang, Jojo Wang, Letian Wang, Shizun Wang, Weizhi Wang, Zixuan Xu, Jinfan Xing, Sen Yang, Chenyu Ye, Hai Yu, Jiaheng Yu, Yue Zhong, Muyan Zhao, Tianchen Zhu, Xizhou Zhou, Yanpeng Zhang, Yifan Zhu, Zhi |
| author_facet | MiroMind Team Bai, Song Bing, Lidong Chen, Carson Chen, Guanzheng Chen, Yuntao Chen, Zhe Chen, Ziyi Dai, Jifeng Dong, Xuan Dou, Wenhan Deng, Yue Fu, Yunjie Ge, Junqi Han, Chenxia Huang, Tammy Huang, Zhenhang Jiao, Jerry Jiang, Shilei Jiao, Tianyu Jian, Xiaoqi Lei, Lei Li, Ruilin Luo, Gen Li, Tiantong Lin, Xiang Liu, Ziyuan Li, Zhiqi Ni, Jie Ren, Qiang Sun, Pax Su, Shiqian Tao, Chenxin Wang, Bin Wang, Wenhai Wang, Haonan Wang, James Wang, Jin Wang, Jojo Wang, Letian Wang, Shizun Wang, Weizhi Wang, Zixuan Xu, Jinfan Xing, Sen Yang, Chenyu Ye, Hai Yu, Jiaheng Yu, Yue Zhong, Muyan Zhao, Tianchen Zhu, Xizhou Zhou, Yanpeng Zhang, Yifan Zhu, Zhi |
| contents | We present MiroThinker v1.0, an open-source research agent designed to advance tool-augmented reasoning and information-seeking capabilities. Unlike previous agents that only scale up model size or context length, MiroThinker explores interaction scaling at the model level, systematically training the model to handle deeper and more frequent agent-environment interactions as a third dimension of performance improvement. Unlike LLM test-time scaling, which operates in isolation and risks degradation with longer reasoning chains, interactive scaling leverages environment feedback and external information acquisition to correct errors and refine trajectories. Through reinforcement learning, the model achieves efficient interaction scaling: with a 256K context window, it can perform up to 600 tool calls per task, enabling sustained multi-turn reasoning and complex real-world research workflows. Across four representative benchmarks-GAIA, HLE, BrowseComp, and BrowseComp-ZH-the 72B variant achieves up to 81.9%, 37.7%, 47.1%, and 55.6% accuracy respectively, surpassing previous open-source agents and approaching commercial counterparts such as GPT-5-high. Our analysis reveals that MiroThinker benefits from interactive scaling consistently: research performance improves predictably as the model engages in deeper and more frequent agent-environment interactions, demonstrating that interaction depth exhibits scaling behaviors analogous to model size and context length. These findings establish interaction scaling as a third critical dimension for building next-generation open research agents, complementing model capacity and context windows. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_11793 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MiroThinker: Pushing the Performance Boundaries of Open-Source Research Agents via Model, Context, and Interactive Scaling MiroMind Team Bai, Song Bing, Lidong Chen, Carson Chen, Guanzheng Chen, Yuntao Chen, Zhe Chen, Ziyi Dai, Jifeng Dong, Xuan Dou, Wenhan Deng, Yue Fu, Yunjie Ge, Junqi Han, Chenxia Huang, Tammy Huang, Zhenhang Jiao, Jerry Jiang, Shilei Jiao, Tianyu Jian, Xiaoqi Lei, Lei Li, Ruilin Luo, Gen Li, Tiantong Lin, Xiang Liu, Ziyuan Li, Zhiqi Ni, Jie Ren, Qiang Sun, Pax Su, Shiqian Tao, Chenxin Wang, Bin Wang, Wenhai Wang, Haonan Wang, James Wang, Jin Wang, Jojo Wang, Letian Wang, Shizun Wang, Weizhi Wang, Zixuan Xu, Jinfan Xing, Sen Yang, Chenyu Ye, Hai Yu, Jiaheng Yu, Yue Zhong, Muyan Zhao, Tianchen Zhu, Xizhou Zhou, Yanpeng Zhang, Yifan Zhu, Zhi Computation and Language We present MiroThinker v1.0, an open-source research agent designed to advance tool-augmented reasoning and information-seeking capabilities. Unlike previous agents that only scale up model size or context length, MiroThinker explores interaction scaling at the model level, systematically training the model to handle deeper and more frequent agent-environment interactions as a third dimension of performance improvement. Unlike LLM test-time scaling, which operates in isolation and risks degradation with longer reasoning chains, interactive scaling leverages environment feedback and external information acquisition to correct errors and refine trajectories. Through reinforcement learning, the model achieves efficient interaction scaling: with a 256K context window, it can perform up to 600 tool calls per task, enabling sustained multi-turn reasoning and complex real-world research workflows. Across four representative benchmarks-GAIA, HLE, BrowseComp, and BrowseComp-ZH-the 72B variant achieves up to 81.9%, 37.7%, 47.1%, and 55.6% accuracy respectively, surpassing previous open-source agents and approaching commercial counterparts such as GPT-5-high. Our analysis reveals that MiroThinker benefits from interactive scaling consistently: research performance improves predictably as the model engages in deeper and more frequent agent-environment interactions, demonstrating that interaction depth exhibits scaling behaviors analogous to model size and context length. These findings establish interaction scaling as a third critical dimension for building next-generation open research agents, complementing model capacity and context windows. |
| title | MiroThinker: Pushing the Performance Boundaries of Open-Source Research Agents via Model, Context, and Interactive Scaling |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2511.11793 |