AIAny
Icon for item

AREX: Towards a Recursively Self-Improving Agent for Deep Research

Alternates targeted research and constraint-wise audits to recursively improve long-horizon answers: an inner loop gathers evidence and drafts solutions, an outer loop audits unresolved claims and launches focused follow-ups. Trains 4B dense and 122B-A10B MoE agents with long-horizon RL and agentic mid-training, outperforming comparable-scale baselines on multi-step research benchmarks.

Introduction

Long research problems are asymmetric: discovering a plausible candidate is expensive, but verifying parts of it can be decomposed into many tractable checks. That asymmetry suggests a different strategy than blind search — repeatedly verify intermediate claims, compress verified evidence, and use the partial verification state to guide focused refinement. AREX operationalizes this intuition into a recursively self-improving research agent.

Key Findings
  • Recursive research loop: AREX alternates an inner research loop (evidence gathering and provisional answer construction) with an outer self-improvement loop (constraint-wise auditing, claim identification, targeted follow-ups). This structure converts sparse final rewards into denser, meaningful learning signals.
  • Learned context-update tool: to sustain long-horizon reasoning, AREX learns an autonomous compressor that updates the agent's improvement state (keeping verified evidence and unresolved constraints) without relying on an external model or unbounded history.
  • Training recipe and architectures: AREX is trained via agentic mid-training and long-horizon reinforcement learning with emphasis on decisive moments (evidence acquisition or correction). The authors instantiate both a dense 4B model and a 122B-A10B Mixture-of-Experts model.
  • Empirical payoff: across BrowseComp, WideSearch, DeepSearchQA and Humanity's Last Exam (HLE), AREX substantially outperforms comparable-scale baselines and competes with models that use many more activated parameters, demonstrating the practical value of recursive verification-guided refinement.
Who it's for and trade-offs

Great fit if you research autonomous agents for long-horizon problem-solving, multi-step tool use, or automated scientific workflows — especially when verification can be factorized into constraint-level checks. Look elsewhere if your tasks lack decomposable verification signals (so recursive auditing offers little guidance), or if you need a lightweight, single-shot solver rather than a long-horizon, RL-trained agent. AREX's benefits come with increased training complexity, the need for well-designed verification checks, and reliance on substantial compute for the larger MoE instantiation.

How it works (concise)
  • Inner loop: uses tool calls and search to collect evidence and assemble a provisional answer.
  • Outer loop: audits the provisional answer constraint-wise, labels unresolved claims, and spawns targeted follow-up research to resolve those claims.
  • Context-update tool: compresses growing interaction history into a compact improvement state that preserves verified evidence and outstanding constraints, enabling sustained recursion without exploding context size.
  • Optimization: combines agentic mid-training trajectories with long-horizon RL and shaped signals focusing on decisive corrective or evidential steps to mitigate reward sparsity.

Taken together, AREX reframes long-horizon research as alternating construction and audit phases, and demonstrates that learned state compression plus targeted follow-ups yield practical gains on multi-step research benchmarks.

Information

  • Websitearxiv.org
  • OrganizationsBeijing Academy of Artificial Intelligence (BAAI)
  • AuthorsShuqi Lu, Chaofan Li, Kun Luo, Zhang Zhang, Hui Wang, Hongwang Xiao, Zheng Liu, Lei Xiong, Jiahao Wang, Sen Wang
  • Published date2026/07/23

Categories

More Items

Guides an LLM agent to build persistent, editable DAG-based data pipelines via typed, incremental mutations instead of free-form scripts. Combines DataFlow-Skills, a Model Context Protocol exposing live operator registry and pipeline state, and a synchronized Web UI; achieves 93.3% end-to-end pass rate on a 12-task benchmark while cutting cost and latency versus script baselines.

Models long-horizon interactive literary simulation where characters and world co-evolve; introduces an open‑schema framework with a Character Agent and an LLM-based World Model, plus seven trainable tasks and a dataset from 57 books for benchmarking persistent narrative state.

Prunes tool-output lines inside a coding LLM agent by turning the agent's own internal representations into per-line keep-or-prune labels. Implements a small classification head plus a length-aware embedding, saving up to 39% of tokens across benchmarks while preserving task quality.