# MuLA-Bench: A Multilingual Long-Form Audio Understanding Benchmark via Multi-Tier Auditing

arXiv: https://arxiv.org/abs/2609.23416 · HTML: https://mlsanity.com/paper/2609.23416/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 11 references that are in the map.
- First posted on arXiv 20 Sep 2026.

## Builds on (most cited first, 11 of 11)

- [MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmark](https://mlsanity.com/paper/2410.19168/) (2024)
- [AIR-Bench: Benchmarking Large Audio-Language Models via Generative Comprehension](https://mlsanity.com/paper/2402.07729/) (2024)
- [AudioBench: A Universal Benchmark for Audio Large Language Models](https://mlsanity.com/paper/2406.16020/) (2024)
- [MMAR: A Challenging Benchmark for Deep Reasoning in Speech, Audio, Music, and Their Mix](https://mlsanity.com/paper/2505.13032/) (2025)
- [Benchmarking Open-ended Audio Dialogue Understanding for Large Audio-Language Models](https://mlsanity.com/paper/2412.05167/) (2024)
- [Listening Between the Frames: Bridging Temporal Gaps in Large Audio-Language Models](https://mlsanity.com/paper/2511.11039/) (2025)
- [ChronosAudio: A Comprehensive Long-Audio Benchmark for Evaluating Audio-Large Language Models](https://mlsanity.com/paper/2601.04876/) (2026)
- [LongSpeech: A Scalable Benchmark for Transcription, Translation and Understanding in Long Speech](https://mlsanity.com/paper/2601.13539/) (2026)
- [SCENEBench: An Audio Understanding Benchmark Grounded in Assistive and Industrial Use Cases](https://mlsanity.com/paper/2603.09853/) (2026)
- [LongAudioSpan: Spanning the Duration and Depth of Audio Comprehension](https://mlsanity.com/paper/2608.26431/) (2026)
- [VoiceGiraffe: A Benchmark for Extreme Long-Context Audio-Language Understanding](https://mlsanity.com/paper/2605.27976/) (2026)

## Built on by (most cited first, 1 of 1)

- [Qwen3.8-Omni: Towards Native Omni-Modal Agents](https://mlsanity.com/paper/2609.25611/) (2026)

## Abstract (arXiv)

Long-form audio performance is often summarized by context length and aggregate accuracy, obscuring how language, evidence, and task jointly shape difficulty. We introduce MuLA-Bench: 5,038 open-ended questions over 1,769 in-the-wild recordings totaling 1,377.9 hours, covering 16 languages and eight domains. A balanced Language x Domain semantic track supports controlled comparisons, while a complementary acoustic track preserves naturally occurring non-speech evidence. Evidence-grounded generation, shortcut checks, and language-expert review provide auditable questions without translating a shared source set or injecting target sounds. We evaluate ten audio-language models and conduct pooled diagnostics on a fixed eight-model cohort. Language rankings change across domains and tasks; acoustic-semantic performance gaps vary with the requested operation; and temporal errors can persist after the correct event is identified. Long-range retrieval is comparatively strong, while precise clock alignment and factual grounding of natural acoustic events remain fragile. MuLA-Bench thus exposes conditional failure patterns that a single long-context score does not capture.