AudioRAG: A Challenging Benchmark for Audio Reasoning and Information Retrieval

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Lin, Jingru, Zhang, Chen, Wang, Tianrui, Li, Haizhou
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910018879619072
author Lin, Jingru
Zhang, Chen
Wang, Tianrui
Li, Haizhou
author_facet Lin, Jingru
Zhang, Chen
Wang, Tianrui
Li, Haizhou
contents Due to recent advancements in Large Audio-Language Models (LALMs) that demonstrate remarkable performance across a range of sound-, speech- and music-related tasks, there is a growing interest in proposing benchmarks to assess these models. Existing benchmarks generally focus only on reasoning with internal knowledge, neglecting real-world scenarios that require external information grounding. To bridge this gap, we introduce AudioRAG, a novel benchmark designed to evaluate audio-based reasoning augmented by information retrieval in realistic web environments. This benchmark comprises both LLM-generated and manually curated question-answer pairs. Our evaluations reveal that even the state-of-the-art LALMs struggle to answer these questions. We therefore propose an agentic pipeline that integrates audio reasoning with retrieval-augmented generation, providing a stronger baseline for future research.
format Preprint
id arxiv_https___arxiv_org_abs_2602_10656
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle AudioRAG: A Challenging Benchmark for Audio Reasoning and Information Retrieval
Lin, Jingru
Zhang, Chen
Wang, Tianrui
Li, Haizhou
Audio and Speech Processing
Sound
Due to recent advancements in Large Audio-Language Models (LALMs) that demonstrate remarkable performance across a range of sound-, speech- and music-related tasks, there is a growing interest in proposing benchmarks to assess these models. Existing benchmarks generally focus only on reasoning with internal knowledge, neglecting real-world scenarios that require external information grounding. To bridge this gap, we introduce AudioRAG, a novel benchmark designed to evaluate audio-based reasoning augmented by information retrieval in realistic web environments. This benchmark comprises both LLM-generated and manually curated question-answer pairs. Our evaluations reveal that even the state-of-the-art LALMs struggle to answer these questions. We therefore propose an agentic pipeline that integrates audio reasoning with retrieval-augmented generation, providing a stronger baseline for future research.
title AudioRAG: A Challenging Benchmark for Audio Reasoning and Information Retrieval
topic Audio and Speech Processing
Sound
url https://arxiv.org/abs/2602.10656