LongHorizon-Harness：推進長程智能體於真實世界任務之應用

摘要

大型語言模型（LLM）智能體日益承擔需要持續推理、工具使用以及跨越多個相互依賴步驟進行修正的長時程任務。然而，現有的智能體框架在不斷增長的上下文中維護任務執行、任務狀態與完成度評估，使得狀態難以追蹤，且不正確的自我評估會傳播至後續決策中。我們將長時程執行重新表述為任務狀態管理問題，並提出LongHorizon-Harness，該框架在執行流程外部明確維護任務狀態，僅以從環境中獨立驗證的事實更新狀態。其管理-執行-審計（MEA）循環使用管理器維護任務狀態並決定下一個子任務，使用全新上下文的執行器執行子任務，並使用唯讀審計器在下一輪之前驗證結果環境狀態。輕量級的AgentAdapter支援可替換的模型與框架後端，無需修改其原生智能體循環。LongHorizon-Harness在WeaveBench上將Qwen 3.7-Plus從51.8%提升至80.7%，在Terminal-Bench 2.1上從69.7%提升至77.2%，在OSWorld 2.0上從2.8%提升至8.3%。它還將Claude Opus 4.7在OSWorld 2.0子集上從20.0%提升至34.3%，證明在模型、框架與互動領域之間的一致提升。

English

Large language model (LLM) agents increasingly undertake long-horizon tasks that require sustained reasoning, tool use, and revision across many interdependent steps. However, existing agent harnesses maintain task execution, task state, and completion assessment within a growing context, making the state difficult to track and allowing incorrect self-assessments to propagate into later decisions. We reformulate long-horizon execution as a task-state management problem and propose LongHorizon-Harness, which maintains the task state explicitly outside execution and updates it only with facts independently verified from the environment. Its Manage-Execute-Audit(MEA) loop uses a manager to maintain the task state and determine the next subtask, a fresh-context executor to perform it, and a read-only auditor to verify the resulting environment state before the next round. A lightweight AgentAdapter supports interchangeable model and harness backends without modifying their native agent loops. LongHorizon-Harness improves Qwen~3.7-Plus from 51.8% to 80.7% on WeaveBench, from 69.7% to 77.2% on Terminal-Bench~2.1, and from 2.8% to 8.3% on OSWorld~2.0. It also raises Claude Opus~4.7 from 20.0% to 34.3% on an OSWorld2.0 subset, demonstrating consistent gains across models, harnesses, and interaction domains.