mlsanityData to 7 Oct 2026

Audio-language model reasoning

303 papers1 this weekLow confidence in this grouping: the papers may belong to more than one topic, or the group may changeLow confidence in this grouping: the papers may belong to more than one topic, or the group may change · About

Keywords: audio, emotion, speech, audio-visual, reasoning

Key facts. Audio-language model reasoning: 303 papers on arXiv since 2022 (data through 7 Oct 2026). 81 of its 303 mapped papers (27%) cite “Qwen2.5-Omni Technical Report” (2025, arXiv 2503.20215). 62 of its 303 mapped papers (20%) cite “MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmark” (2024, arXiv 2410.19168). 21 new papers 17 Sep–7 Oct.

What’s new

17 Sep – 7 Oct 2026

21 new papers, 17 Sep – 7 Oct; 1 this week.

  1. This week · 1
  2. Mind the Accent Gap: British Accent Robustness in Speech-Driven Financial Voice Assistants
  3. Earlier · 20
  4. AdaLoop: Adaptive-Depth Latent Reasoning for Audio Language Models
  5. OmniSeek: Native Tool Integration for Multi-turn Audio-Visual Reasoning also fits: Multimodal reasoning
  6. AURAL: Adaptive Latent Reasoning with Joint Chunk for Speech Language Models
  7. Align Then Reason: A Multimodal Lip-Sync Judge for Dubbing
  8. LongEmo: Towards Emotion Understanding and Reasoning in Long Videos
  9. Learning to Reason with Compressed Context: Ground-Truth-Free Adaptation of OmniLLMs via Self-Distillation
  10. SEAR: Spoofing Evidence-Grounded Audio Reasoning Benchmark for Audio Language Models
  11. OP-CAD: On-Policy Clean-Audio Distillation for Robust Audio-Visual Reasoning
  12. OmniRoute: Mapping Temporal Semantic Evidence to Audio-Visual Token Budgets for Efficient Omnimodal Large Language Models
All 21
  1. Decoding Affective Nuances: Enhancing MLLMs via Hierarchical Emotion Reasoning and Contrastive Discriminative Pruning
  2. Emo-Jev: Probabilistic Reasoning for Emotion Classification with Jev
  3. DuraS2ST: Chain-of-Thought and Reinforcement Learning for Duration-Aligned Speech-to-Speech Translation
  4. OmniSmartHome: A Multimodal Reasoning Benchmark for Smart-Home Agents also fits: Spatial and embodied reasoning
  5. Who Says What: Symbolic Trimodal Binding Mechanisms in Audio-Visual LLMs
  6. Rethinking Length-Based Training: Batch Composition and Loss Normalization in Speech Token Language Models
  7. Listen Then Reason: Perception-Grounded Test-Time Reinforcement Learning for Large Audio-Language Models
  8. COT-TTS: Audio Context-Aware Text-to-Speech with Chain-of-Thought Reasoning
  9. SEABED: SouthEast Asian Benchmark for Evaluating Audio Reasoning
  10. Rethinking Human-Aligned Evaluation: An Analysis of Semantic Metrics Beyond WER
  11. AVTrace: Diagnosing Audio-Visual Temporal Reasoning in Omni Models also fits: Multimodal reasoning

Foundations

Most cited by papers here

  1. Qwen2.5-Omni Technical Report2025cited by 81 here
  2. MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmark2024cited by 62 here
  3. Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language Models2025cited by 46 here
  4. Audio-Reasoner: Improving Reasoning Capability in Large Audio Language Models2025cited by 43 here
  5. MMAR: A Challenging Benchmark for Deep Reasoning in Speech, Audio, Music, and Their Mix2025cited by 36 here

Lineage

The papers this direction builds on, oldest first.

Each paper on the line cites the one before.

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 86 here
  2. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 22 here
  3. 2022Large Language Models are Zero-Shot Reasonerscited by 18 here
  4. 2024GAMA: A Large Audio-Language Model with Advanced Audio Understanding and Complex Reasoning Abilitiescited by 32 here
  5. 2024MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmarkcited by 62 here
  6. 2025Audio-CoT: Exploring Chain-of-Thought Reasoning in Large Audio Language Modelcited by 27 here
  7. 2025Audio-Reasoner: Improving Reasoning Capability in Large Audio Language Modelscited by 43 here
  8. 2025SARI: Structured Audio Reasoning via Curriculum-Guided Reinforcement Learningcited by 16 here
  9. 2025MMAR: A Challenging Benchmark for Deep Reasoning in Speech, Audio, Music, and Their Mixcited by 36 here
  10. 2025Audio-Thinker: Guiding Audio Language Model When and How to Think via Reinforcement Learningcited by 14 here
14 more it builds on, not in this chain
  1. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 30 here
  2. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 15 here
  3. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 20 here
  4. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 49 here
  5. 2024Emotion-LLaMA: Multimodal Emotion Recognition and Reasoning with Instruction Tuningcited by 25 here
  6. 2024Mini-Omni: Language Models Can Hear, Talk While Thinking in Streamingcited by 17 here
  7. 2024OpenAI o1 System Cardcited by 15 here
  8. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 59 here
  9. 2025Audio Flamingo 2: An Audio-Language Model with Long-Audio Understanding and Expert Reasoning Abilitiescited by 29 here
  10. 2025Reinforcement Learning Outperforms Supervised Fine-Tuning: A Case Study on Audio Question Answeringcited by 28 here
  11. 2025Qwen2.5-Omni Technical Reportcited by 81 here
  12. 2025Qwen3 Technical Reportcited by 44 here
  13. 2025MMSU: A Massive Multi-task Spoken Language Understanding and Reasoning Benchmarkcited by 16 here
  14. 2025Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language Modelscited by 46 here

The 24 most cited of 666.

Browse

30 of 303, ranked 6–35

  1. 6GAMA: A Large Audio-Language Model with Advanced Audio Understanding and Complex Reasoning Abilities202432
  2. 7Audio Flamingo 2: An Audio-Language Model with Long-Audio Understanding and Expert Reasoning Abilities202529
  3. 8Reinforcement Learning Outperforms Supervised Fine-Tuning: A Case Study on Audio Question Answering202528
  4. 9Audio-CoT: Exploring Chain-of-Thought Reasoning in Large Audio Language Model202527
  5. 10Emotion-LLaMA: Multimodal Emotion Recognition and Reasoning with Instruction Tuning202425
  6. 11Mini-Omni: Language Models Can Hear, Talk While Thinking in Streaming202417
  7. 12SARI: Structured Audio Reasoning via Curriculum-Guided Reinforcement Learning202516
  8. 13MMSU: A Massive Multi-task Spoken Language Understanding and Reasoning Benchmark202516
  9. 14Audio-Thinker: Guiding Audio Language Model When and How to Think via Reinforcement Learning202514
  10. 15Daily-Omni: Towards Audio-Visual Reasoning with Temporal Alignment across Modalities202513
All 30
  1. 16Step-Audio-R1 Technical Report202512
  2. 17EchoInk-R1: Exploring Audio-Visual Reasoning in Multimodal LLMs via Reinforcement Learning202511
  3. 18Mellow: a small audio language model for reasoning202510
  4. 19SAKURA: On the Multi-hop Reasoning of Large Audio-Language Models Based on Speech and Audio Information20259
  5. 20OmniVideoBench: Towards Audio-Visual Understanding Evaluation for Omni MLLMs20259
  6. 21BAT: Learning to Reason about Spatial Sounds with Large Language Models20249
  7. 22Benchmarking and Bridging Emotion Conflicts for Multimodal Emotion Reasoning20258
  8. 23SoundMind: RL-Incentivized Logic Reasoning for Audio-Language Models20257
  9. 24CompA: Addressing the Gap in Compositional Reasoning in Audio-Language Models20237
  10. 25video-SALMONN-o1: Reasoning-enhanced Audio-visual Large Language Model20256
  11. 26STITCH: Simultaneous Thinking and Talking with Chunked Reasoning for Spoken Language Models20256
  12. 27SpeechR: A Benchmark for Speech Reasoning in Large Audio-Language Models20256
  13. 28Mini-Omni-Reasoner: Token-Level Thinking-in-Speaking in Large Speech Models20256
  14. 29Crab: A Unified Audio-Visual Scene Understanding Model with Explicit Cooperation20255
  15. 30Aurelia: Test-time Reasoning Distillation in Audio-Visual LLMs20255
  16. 31Incentivizing Consistent, Effective and Scalable Reasoning Capability in Audio LLMs via Reasoning Process Rewards20255
  17. 32ExpLLM: Towards Chain of Thought for Facial Expression Recognition20245
  18. 33FEALLM: Advancing Facial Emotion Analysis in Multimodal Large Language Models with Emotional Synergy and Reasoning20254
  19. 34OWL: Geometry-Aware Spatial Reasoning for Audio Large Language Models20254
  20. 35Can Speech LLMs Think while Listening?20254