MAXS: Meta-Adaptive Exploration with LLM Agents

TL;DR

MAXS introduces a meta-adaptive reasoning framework for LLM agents that employs lookahead and trajectory convergence to improve reasoning stability, efficiency, and overall performance across multiple models and datasets.

Contribution

The paper presents MAXS, a novel meta-adaptive exploration method that integrates lookahead planning and trajectory convergence to enhance LLM agent reasoning and efficiency.

Findings

MAXS outperforms existing methods in accuracy and efficiency.

Lookahead strategy improves reasoning stability.

Trajectory convergence reduces computational costs.

Abstract

Large Language Model (LLM) Agents exhibit inherent reasoning abilities through the collaboration of multiple tools. However, during agent inference, existing methods often suffer from (i) locally myopic generation, due to the absence of lookahead, and (ii) trajectory instability, where minor early errors can escalate into divergent reasoning paths. These issues make it difficult to balance global effectiveness and computational efficiency. To address these two issues, we propose meta-adaptive exploration with LLM agents https://github.com/exoskeletonzj/MAXS, a meta-adaptive reasoning framework based on LLM Agents that flexibly integrates tool execution and reasoning planning. MAXS employs a lookahead strategy to extend reasoning paths a few steps ahead, estimating the advantage value of tool usage, and combines step consistency variance and inter-step trend slopes to jointly select stable, consistent, and high-value reasoning steps. Additionally, we introduce a trajectory convergence mechanism that controls computational cost by halting further rollouts once path consistency is achieved, enabling a balance between resource efficiency and global effectiveness in multi-tool reasoning. We conduct extensive empirical studies across three base models (MiMo-VL-7B, Qwen2.5-VL-7B, Qwen2.5-VL-32B) and five datasets, demonstrating that MAXS consistently outperforms existing methods in both performance and inference efficiency. Further analysis confirms the effectiveness of our lookahead strategy and tool usage.