A$^2$FM: An Adaptive Agent Foundation Model for Tool-Aware Hybrid Reasoning
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912662285189120 |
|---|---|
| author | Chen, Qianben Cao, Jingyi Zhang, Jiayu Qin, Tianrui Li, Xiaowan Zhu, King Shi, Dingfeng Zhu, He Liu, Minghao Liang, Xiaobo Gui, Xin Zhang, Ge Yang, Jian Jiang, Yuchen Eleanor Zhou, Wangchunshu |
| author_facet | Chen, Qianben Cao, Jingyi Zhang, Jiayu Qin, Tianrui Li, Xiaowan Zhu, King Shi, Dingfeng Zhu, He Liu, Minghao Liang, Xiaobo Gui, Xin Zhang, Ge Yang, Jian Jiang, Yuchen Eleanor Zhou, Wangchunshu |
| contents | Large language models split into two families: reasoning-centric LLMs, which strengthen internal chain-of-thought reasoning but cannot invoke external tools, and agentic LLMs, which learn to interact with environments and leverage tools but often lag in deep reasoning. This divide arises from fundamentally different training objectives, leading to mismatched strengths and inefficiency on simple queries, where both families tend to overthink or over-call tools. In this work, we present Adaptive Agent Foundation Model (A$^2$FM), a unified framework that follows a route-then-align principle: the model first learns task-aware routing and then aligns mode-specific trajectories under a shared backbone. To address the inefficiency gap, we introduce a third mode-instant-that handles simple queries directly, preventing unnecessary reasoning or tool calls while complementing the agentic and reasoning modes. To jointly enhance accuracy and efficiency, we propose Adaptive Policy Optimization (APO), which enforces adaptive sampling across modes and applies a cost-regularized reward. On the 32B scale, A$^2$FM achieves 13.4% on BrowseComp, 70.4% on AIME25, and 16.7% on HLE, setting new SOTA among comparable models and performing competitively with frontier LLMs across agentic, reasoning, and general benchmarks. Notably, the adaptive execution achieves a cost of pass of only $0.00487 per correct answer-cutting cost by 45.2% relative to reasoning and 33.5% relative to agentic, thus delivering substantially higher cost efficiency while maintaining comparable accuracy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_12838 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A$^2$FM: An Adaptive Agent Foundation Model for Tool-Aware Hybrid Reasoning Chen, Qianben Cao, Jingyi Zhang, Jiayu Qin, Tianrui Li, Xiaowan Zhu, King Shi, Dingfeng Zhu, He Liu, Minghao Liang, Xiaobo Gui, Xin Zhang, Ge Yang, Jian Jiang, Yuchen Eleanor Zhou, Wangchunshu Computation and Language Artificial Intelligence Large language models split into two families: reasoning-centric LLMs, which strengthen internal chain-of-thought reasoning but cannot invoke external tools, and agentic LLMs, which learn to interact with environments and leverage tools but often lag in deep reasoning. This divide arises from fundamentally different training objectives, leading to mismatched strengths and inefficiency on simple queries, where both families tend to overthink or over-call tools. In this work, we present Adaptive Agent Foundation Model (A$^2$FM), a unified framework that follows a route-then-align principle: the model first learns task-aware routing and then aligns mode-specific trajectories under a shared backbone. To address the inefficiency gap, we introduce a third mode-instant-that handles simple queries directly, preventing unnecessary reasoning or tool calls while complementing the agentic and reasoning modes. To jointly enhance accuracy and efficiency, we propose Adaptive Policy Optimization (APO), which enforces adaptive sampling across modes and applies a cost-regularized reward. On the 32B scale, A$^2$FM achieves 13.4% on BrowseComp, 70.4% on AIME25, and 16.7% on HLE, setting new SOTA among comparable models and performing competitively with frontier LLMs across agentic, reasoning, and general benchmarks. Notably, the adaptive execution achieves a cost of pass of only $0.00487 per correct answer-cutting cost by 45.2% relative to reasoning and 33.5% relative to agentic, thus delivering substantially higher cost efficiency while maintaining comparable accuracy. |
| title | A$^2$FM: An Adaptive Agent Foundation Model for Tool-Aware Hybrid Reasoning |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2510.12838 |