mlsanityData to 7 Oct 2026

SEA-SpeechBench: A Large-Scale Multitask Benchmark for Speech Understanding Across Southeast Asia

arXiv 2609.09672 ↗First posted 9 Sep 2026cs.CLcited by 0

On the map: Multimodal LLMs / vision-language models → MLLM training and visual instruction tuning

Abstract (arXiv)

The rapid advancement of audio and multimodal large language models has unlocked transformative speech understanding capabilities, yet evaluation frameworks remain predominantly English-centric, leaving Southeast Asian (SEA) languages critically underrepresented. We introduce SEA-SpeechBench, to the best of our knowledge, the first large-scale multitask benchmark that evaluates speech understanding in 11 SEA languages through 97,194 samples across 99 evaluation sets and 597 hours of curated audio data. Our benchmark comprises 9 diverse tasks across 3 categories: speech processing (automatic speech recognition, speech translation, spoken question answering), paralinguistic analysis (emotion, gender, age, speaker recognition), and temporal understanding, a novel dimension featuring timestamped content queries and temporal localization within extended audio sequences up to 3 minutes. We implement multilingual prompting in both native SEA languages and English to reflect user interactions with audio-language models. Evaluation of leading open-source and proprietary systems reveals marked performance gaps. Across all models, performance remains underwhelming on temporal understanding, emotion recognition, and speech translation. Prompting in low-resource languages such as Burmese and Tamil lags behind English by up to 41 percentage points. Our findings expose critical model limitations and underscore the need for inclusive model development. The SEA-SpeechBench benchmark is available at https://zwenyu.github.io/SEA-SpeechBench/.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in MLLM training and visual instruction tuning, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 20 of its 43 references that are in the map. First posted on arXiv 9 Sep 2026.

Builds on

20 of 43 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025Phi-4-Mini Technical Report: Compact yet Powerful Multimodal Language Models via Mixture-of-LoRAs3 Mar
  2. 2023NExT-GPT: Any-to-Any Multimodal LLM11 Sep
  3. 2023SpeechGPT: Empowering Large Language Models with Intrinsic Cross-Modal Conversational Abilities18 May

Other references · 17

  1. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 3,972
  2. 2019Common Voice: A Massively-Multilingual Speech Corpuscited by 1,358
  3. 2021SUPERB: Speech processing Universal PERformance Benchmarkcited by 652
  4. 2022FLEURS: Few-shot Learning Evaluation of Universal Representations of Speechcited by 458
  5. 2021VoxPopuli: A Large-Scale Multilingual Speech Corpus for Representation Learning, Semi-Supervised Learning and Interpretationcited by 455
  6. 2017AISHELL-1: An Open-Source Mandarin Speech Corpus and A Speech Recognition Baselinecited by 364
  7. 2024MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmarkcited by 281
  8. 2023On decoder-only architecture for speech-to-text and large language model integrationcited by 144
  9. 2024YODAS: Youtube-Oriented Dataset for Audio and Speechcited by 56
  10. 2022M3ED: Multi-modal Multi-scene Multi-label Emotional Dialogue Databasecited by 50
  11. 2022Bloom Library: Multimodal Datasets in 300+ Languages for a Variety of Downstream Taskscited by 28
  12. 2025CMI-Bench: A Comprehensive Benchmark for Evaluating Music Instruction Followingcited by 15
5 more
  1. 2024VietMed: A Dataset and Benchmark for Automatic Speech Recognition of Vietnamese in the Medical Domaincited by 9
  2. 2019Enriching Rare Word Representations in Neural Language Models by Embedding Matrix Augmentationcited by 7
  3. 2025SeaLLMs-Audio: Large Audio-Language Models for Southeast Asiacited by 7
  4. 2024VoxVietnam: a Large-Scale Multi-Genre Dataset for Vietnamese Speaker Recognitioncited by 3
  5. 2025THAI Speech Emotion Recognition (THAI-SER) corpuscited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026SEABED: SouthEast Asian Benchmark for Evaluating Audio Reasoning
  2. 2026SEA-NLI: Natural Language Inference as a Lens into Southeast Asian Cultural Understanding
  3. 2025MMSU: A Massive Multi-task Spoken Language Understanding and Reasoning Benchmark
  4. 2026PolySpeech-100: A Large-Scale Benchmark for Speech Understanding Across 100+ Languages and Dialects
  5. 2025Advancing Singlish Understanding: Bridging the Gap with Datasets and Multimodal Models
  6. 2026SONIC-O1: A Real-World Benchmark for Evaluating Multimodal Large Language Models on Audio-Video Understanding
  7. 2024MERaLiON-SpeechEncoder: Towards a Speech Foundation Model for Singapore and Beyond
  8. 2026Efficiently Adapting Spoken Language Models for the Singaporean Context