InfiAgent: An Infinite-Horizon Framework for General-Purpose Autonomous Agents
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915711393202176 |
|---|---|
| author | Yu, Chenglin Wang, Yuchen Wang, Songmiao Yang, Hongxia Li, Ming |
| author_facet | Yu, Chenglin Wang, Yuchen Wang, Songmiao Yang, Hongxia Li, Ming |
| contents | LLM agents can reason and use tools, but they often break down on long-horizon tasks due to unbounded context growth and accumulated errors. Common remedies such as context compression or retrieval-augmented prompting introduce trade-offs between information fidelity and reasoning stability. We present InfiAgent, a general-purpose framework that keeps the agent's reasoning context strictly bounded regardless of task duration by externalizing persistent state into a file-centric state abstraction. At each step, the agent reconstructs context from a workspace state snapshot plus a fixed window of recent actions. Experiments on DeepResearch and an 80-paper literature review task show that, without task-specific fine-tuning, InfiAgent with a 20B open-source model is competitive with larger proprietary systems and maintains substantially higher long-horizon coverage than context-centric baselines. These results support explicit state externalization as a practical foundation for stable long-horizon agents. Github Repo:https://github.com/ChenglinPoly/infiAgent |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_03204 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | InfiAgent: An Infinite-Horizon Framework for General-Purpose Autonomous Agents Yu, Chenglin Wang, Yuchen Wang, Songmiao Yang, Hongxia Li, Ming Artificial Intelligence Multiagent Systems LLM agents can reason and use tools, but they often break down on long-horizon tasks due to unbounded context growth and accumulated errors. Common remedies such as context compression or retrieval-augmented prompting introduce trade-offs between information fidelity and reasoning stability. We present InfiAgent, a general-purpose framework that keeps the agent's reasoning context strictly bounded regardless of task duration by externalizing persistent state into a file-centric state abstraction. At each step, the agent reconstructs context from a workspace state snapshot plus a fixed window of recent actions. Experiments on DeepResearch and an 80-paper literature review task show that, without task-specific fine-tuning, InfiAgent with a 20B open-source model is competitive with larger proprietary systems and maintains substantially higher long-horizon coverage than context-centric baselines. These results support explicit state externalization as a practical foundation for stable long-horizon agents. Github Repo:https://github.com/ChenglinPoly/infiAgent |
| title | InfiAgent: An Infinite-Horizon Framework for General-Purpose Autonomous Agents |
| topic | Artificial Intelligence Multiagent Systems |
| url | https://arxiv.org/abs/2601.03204 |