The Sonar Moment: Benchmarking Audio-Language Models in Audio Geo-Localization

TL;DR

This paper introduces AGL1K, a new benchmark dataset for audio geo-localization, evaluates existing models, and analyzes their reasoning, biases, and interpretability to advance audio-language model capabilities in geographic inference.

Contribution

The paper presents the first high-quality audio geo-localization benchmark, AGL1K, and provides comprehensive evaluation and analysis of audio-language models' localization abilities.

Findings

Closed-source models outperform open-source models.

Linguistic clues are often primary for localization.

The benchmark enables future advancements in geospatial reasoning.

Abstract

Geo-localization aims to infer the geographic origin of a given signal. In computer vision, geo-localization has served as a demanding benchmark for compositional reasoning and is relevant to public safety. In contrast, progress on audio geo-localization has been constrained by the lack of high-quality audio-location pairs. To address this gap, we introduce AGL1K, the first audio geo-localization benchmark for audio language models (ALMs), spanning 72 countries and territories. To extract reliably localizable samples from a crowd-sourced platform, we propose the Audio Localizability metric that quantifies the informativeness of each recording, yielding 1,444 curated audio clips. Evaluations on 16 ALMs show that ALMs have emerged with audio geo-localization capability. We find that closed-source models substantially outperform open-source models, and that linguistic clues often dominate as a scaffold for prediction. We further analyze ALMs' reasoning traces, regional bias, error causes, and the interpretability of the localizability metric. Overall, AGL1K establishes a benchmark for audio geo-localization and may advance ALMs with better geospatial reasoning capability.