GenericAgent: A Token-Efficient Self-Evolving LLM Agent via Contextual Information Density Maximization (V1.0)

2026-04-18

AI Summary

This work addresses the challenge that long-horizon large language model (LLM) agents, constrained by limited context windows, struggle to retain critical decision-making information and reuse cross-task experience over extended interactions. To overcome this, the authors propose a principle of maximizing contextual information density and introduce a general, self-evolving agent framework. This framework integrates a minimal atomic toolset, hierarchical on-demand memory retrieval, trajectory-driven executable standard operating procedure (SOP) self-evolution, and context compression strategies to efficiently preserve high-value information within tight context budgets. Experiments demonstrate that the proposed system significantly outperforms existing approaches in task completion rate, tool usage efficiency, memory effectiveness, self-evolution capability, and web navigation performance, while substantially reducing token consumption and interaction steps, thereby enabling continuous autonomous improvement.

Technology Category

• Multiagent Systems: Agent Communication
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Long-horizon large language model (LLM) agents are fundamentally limited by context. As interactions become longer, tool descriptions, retrieved memories, and raw environmental feedback accumulate and push out the information needed for decision-making. At the same time, useful experience gained from tasks is often lost across episodes. We argue that long-horizon performance is determined not by context length, but by how much decision-relevant information is maintained within a finite context budget. We present GenericAgent (GA), a general-purpose, self-evolving LLM agent system built around a single principle: context information density maximization. GA implements this through four closely connected components: a minimal atomic tool set that keeps the interface simple, a hierarchical on-demand memory that only shows a small high-level view by default, a self-evolution mechanism that turns verified past trajectories into reusable SOPs and executable code, and a context truncation and compression layer that maintains information density during long executions. Across task completion, tool use efficiency, memory effectiveness, self-evolution, and web browsing, GA consistently outperforms leading agent systems while using significantly fewer tokens and interactions, and it continues to evolve over time. Project: https://github.com/lsdefine/GenericAgent

Problem

• long-horizon LLM agents
• context limitation
• information density
• memory retention
• decision-relevant information

Innovation

• context information density
• self-evolving agent
• hierarchical memory
• atomic tool set
• context compression