Eureka-Audio: Triggering Audio Intelligence in Compact Language Models

TL;DR

Eureka-Audio is a compact, high-performance audio language model with only 1.7B parameters that rivals much larger models across various audio understanding tasks by leveraging a unified architecture and innovative data synthesis techniques.

Contribution

The paper introduces Eureka-Audio, a novel lightweight audio language model with a unique architecture and a data synthesis pipeline, achieving competitive performance with significantly fewer parameters.

Findings

Matches or surpasses larger models on multiple benchmarks.

Demonstrates strong performance in ASR, audio understanding, and captioning.

Balances computational efficiency with high accuracy.

Abstract

We present Eureka-Audio, a compact yet high-performance audio language model that achieves competitive performance against models that are 4 to 18 times larger across a broad range of audio understanding benchmarks. Despite containing only 1.7B parameters, Eureka-Audio demonstrates strong performance on automatic speech recognition (ASR), audio understanding, and dense audio captioning, matching or surpassing multiple 7B to 30B audio and omni-modal baselines. The model adopts a unified end-to-end architecture composed of a lightweight language backbone, a Whisper-based audio encoder, and a sparsely activated Mixture-of-Experts (MoE) adapter that explicitly accounts for audio heterogeneity and alleviates cross-modal optimization conflicts under limited capacity. To further enhance paralinguistic reasoning, we introduce DataFlux, a closed loop audio instruction data synthesis and verification pipeline that constructs high quality, logically consistent supervision from raw audio. Extensive evaluations across ASR, knowledge reasoning, safety, instruction following, and paralinguistic benchmarks, demonstrate that Eureka-Audio achieves an efficient balance between computational cost and performance. These results establish Eureka Audio as a strong and practical baseline for lightweight audio understanding models.