擴展視野，而非參數：以35B智能體達到萬億參數性能

Scaling the Horizon, Not the Parameters: Reaching Trillion-Parameter Performance with a 35B Agent

June 29, 2026

作者: Lei Bai, Zongsheng Cao, Yang Chen, Zhiyao Cui, Shangheng Du, Yue Fan, Shiyang Feng, Zijie Guo, Haonan He, Liang He, Xiaohan He, Shuyue Hu, Yusong Hu, Songtao Huang, Yichen Jiang, Hao Li, Xin Li, Dahua Lin, Weihao Lin, Fenghua Ling, Dongrui Liu, Zhuo Liu, Runmin Ma, Chunjiang Mu, Haoyang Peng, Tianshuo Peng, Jinxin Shi, Luohe Shi, Boyuan Sun, Zelin Tan, Shengji Tang, Qianyi Wang, Yiming Wu, Yi Xie, Xiangchao Yan, Jingqi Ye, Peng Ye, Fangchen Yu, Jiakang Yuan, Bihao Zhan, Bo Zhang, Chen Zhang, Shufei Zhang, Shuaiyu Zhang, Wenlong Zhang, Yiqun Zhang, Junpeng Zhao, Zhijie Zhong, Bowen Zhou, Yuhao Zhou

cs.AI

摘要

我們介紹Agents-A1，這是一個350億參數的混合專家代理模型，透過擴展代理視野達到了萬億參數等級的效能。我們從兩個角度探討代理視野擴展：擴展長視野軌跡與擴展異構代理能力。為支持此目標，我們建構了一個長視野知識-行動基礎設施，連結外部知識、行動、觀察結果與驗證器輸出，產生了平均長度為45K token的代理軌跡。在此基礎上，我們採用三階段訓練方案來訓練Agents-A1。首先，進行全領域監督式微調，使基礎模型與廣泛的代理行為對齊。其次，訓練領域級教師模型以捕捉各領域的專業知識。第三，我們提出一種多教師領域路由的線上蒸餾方法，結合顯著詞彙對齊，以提升跨領域的知識遷移效率，將六個異構領域統合為一個可部署的學生模型。Agents-A1在長視野代理基準測試中展現出強大且廣泛的效能。與Kimi-K2.6和DeepSeek-V4-pro等萬億參數模型相比，Agents-A1在SEAL-0（56.4）、IFBench（80.6）、HiPhO（46.4）、FrontierScience-Olympiad（79.0）及MolBench-Bind（56.8）上取得領先結果，並在SciCode（44.3）、HLE（47.6）及BrowseComp（75.5）上保持高度競爭力。我們希望這項工作能為學界提供一條實用路徑，透過一個350億參數的代理來擴展視野，使其在長視野任務上達到或匹配1萬億參數模型的效能。

English

We introduce Agents-A1, a 35B Mixture-of-Experts Agentic Model that reaches trillion-parameter-level performance by scaling the agent horizon. We investigate agent-horizon scaling from two perspectives: scaling long-horizon trajectories and scaling heterogeneous agent abilities. To support this goal, we build a long-horizon knowledge-action infrastructure that connects external knowledge, actions, observations, and verifier outcomes, producing agentic trajectories with an average length of 45K tokens. Based on this, we train Agents-A1 with a three-stage recipe. First, we perform full-domain supervised fine-tuning to align the base model with broad agentic behaviors. Second, we train domain-level teacher models to capture specialized expertise in each domain. Third, we propose a multi-teacher domain-routed on-policy distillation with salient vocabulary alignment to improve knowledge transfer efficiency across different domains, unifying six heterogeneous domains into one deployable student model. Agents-A1 achieves strong and broad performance for long-horizon agent benchmarks. Compared with 1T-parameter model such as Kimi-K2.6 and DeepSeek-V4-pro, Agents-A1 achieves leading results on SEAL-0 (56.4), IFBench (80.6), HiPhO (46.4), FrontierScience-Olympiad (79.0), and MolBench-Bind (56.8), and remains highly competitive on SciCode (44.3), HLE (47.6) and BrowseComp (75.5). We hope this work provides the community with a practical path for scaling the horizon using a 35B agent that can reach or match the performance of 1T models on long-horizon tasks.