Saved in:
Bibliographic Details
Main Author: Dobrovolskyi, Ivan
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.22367
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914414777597952
author Dobrovolskyi, Ivan
author_facet Dobrovolskyi, Ivan
contents Large Language Models (LLMs) deployed as autonomous agents commonly use Retrieval-Augmented Generation (RAG), feeding retrieved documents into the context window, which creates two problems: the risk of hallucination grows with context length, and token cost scales linearly with dataset size. We propose the Reasoner-Executor-Synthesizer (RES) architecture, a three-layer design that strictly separates intent parsing (Reasoner), deterministic data retrieval and aggregation (Executor), and narrative generation (Synthesizer). The Executor uses zero LLM tokens and passes only fixed-size statistical summaries to the Synthesizer. We formally prove that RES achieves O(1) token complexity with respect to dataset size, and validate this on ScholarSearch, a scholarly research assistant backed by the Crossref API (130M+ articles). Across 100 benchmark runs, RES achieves a mean token cost of 1,574 tokens regardless of whether the dataset contains 42,000 or 16.3 million articles. The architecture eliminates data hallucination by construction: the LLM never sees raw records. KEYWORDS LLM agents; agentic architecture; hallucination elimination; token optimization; context window; retrieval-augmented generation; deterministic execution; scholarly metadata; Crossref API; O(1) complexity.
format Preprint
id arxiv_https___arxiv_org_abs_2603_22367
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Reasoner-Executor-Synthesizer: Scalable Agentic Architecture with Static O(1) Context Window
Dobrovolskyi, Ivan
Information Retrieval
Artificial Intelligence
Large Language Models (LLMs) deployed as autonomous agents commonly use Retrieval-Augmented Generation (RAG), feeding retrieved documents into the context window, which creates two problems: the risk of hallucination grows with context length, and token cost scales linearly with dataset size. We propose the Reasoner-Executor-Synthesizer (RES) architecture, a three-layer design that strictly separates intent parsing (Reasoner), deterministic data retrieval and aggregation (Executor), and narrative generation (Synthesizer). The Executor uses zero LLM tokens and passes only fixed-size statistical summaries to the Synthesizer. We formally prove that RES achieves O(1) token complexity with respect to dataset size, and validate this on ScholarSearch, a scholarly research assistant backed by the Crossref API (130M+ articles). Across 100 benchmark runs, RES achieves a mean token cost of 1,574 tokens regardless of whether the dataset contains 42,000 or 16.3 million articles. The architecture eliminates data hallucination by construction: the LLM never sees raw records. KEYWORDS LLM agents; agentic architecture; hallucination elimination; token optimization; context window; retrieval-augmented generation; deterministic execution; scholarly metadata; Crossref API; O(1) complexity.
title Reasoner-Executor-Synthesizer: Scalable Agentic Architecture with Static O(1) Context Window
topic Information Retrieval
Artificial Intelligence
url https://arxiv.org/abs/2603.22367