---
type: "raw_signal"
id: "baai-released-arex-a-recursively-self-improving-research-eb32c7"
slug: "baai-released-arex-a-recursively-self-improving-research-eb32c7"
title: "BAAI released AREX , a recursively self-improving research agent with inner and outer loops."
description: "BAAI’s AREX paper presents a deep-research agent organized around two loops: an inner loop gathers evidence and constructs a provisional answer, while an outer loop audits claims against individual constraints and launches targeted follow-up research for unresolved points."
retrieval_nugget: "BAAI’s AREX paper presents a deep-research agent organized around two loops: an inner loop gathers evidence and constructs a provisional answer, while an outer loop audits claims against individual constraints and launches targeted follow-up research for unresolved points."
observed_at: "2026-10-11"
record_date: "2026-10-11"
date_kind: "observed_at"
why_it_matters: "Source: https://arxiv.org/abs/2607.21461"
novelty: "notable"
verification_level: "source-linked"
signal_type: "release"
source_platform: "arxiv.org"
topics: []
entities: ["BAAI","AREX","AREX-Turbo","AREX-Base","Qwen3.5","BrowseComp","Humanity's Last Exam"]
related_patterns: ["scientific-agents-need-artifact-evidence","goal-scoped-loops-replace-manual-continuation"]
source_url: "https://arxiv.org/abs/2607.21461"
source_urls: ["https://arxiv.org/abs/2607.21461"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "signal:baai-released-arex-a-recursively-self-improving-research-eb32c7"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/signals/baai-released-arex-a-recursively-self-improving-research-eb32c7/","markdown":"https://newruntime.com/signals/baai-released-arex-a-recursively-self-improving-research-eb32c7.md","json":"https://newruntime.com/signals/baai-released-arex-a-recursively-self-improving-research-eb32c7.json"}
---

# BAAI released AREX , a recursively self-improving research agent with inner and outer loops.

## Retrieval answer

BAAI’s AREX paper presents a deep-research agent organized around two loops: an inner loop gathers evidence and constructs a provisional answer, while an outer loop audits claims against individual constraints and launches targeted follow-up research for unresolved points.

BAAI’s AREX paper presents a deep-research agent organized around two loops: an inner loop gathers evidence and constructs a provisional answer, while an outer loop audits claims against individual constraints and launches targeted follow-up research for unresolved points. The authors frame this design around a discovery–verification asymmetry: finding a candidate that satisfies multiple constraints can be costly, while checking a candidate can often be decomposed into tractable constraint-wise checks.

The paper also describes an autonomous context-update tool that compresses interaction history while retaining verified evidence and unresolved constraints, without relying on an external model. BAAI reports a dense 4B model and a 122B-A10B Mixture-of-Experts model, and reports performance across BrowseComp, WideSearch, DeepSearchQA, Humanity’s Last Exam, and other reasoning and tool-use benchmarks. These are claims reported in the paper; the supplied source does not establish independent replication or deployment outcomes.

Source: https://arxiv.org/abs/2607.21461

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- collection: [Raw Signals](https://newruntime.com/signals.json): Open every source-linked raw signal.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)