mlsanityData to 19 Oct 2026

Audio language models and spoken dialogue

739 papers1 this weekMedium confidence in this grouping: mostly one topic, with blurry edgesMedium confidence in this grouping: mostly one topic, with blurry edges · About

Keywords: audio, reasoning, dialogue, spoken, speech

Key facts. Audio language models and spoken dialogue: 739 papers on arXiv since 2021 (data through 19 Oct 2026). 200 of its 739 mapped papers (27%) cite “Qwen2-Audio Technical Report” (2024, arXiv 2407.10759). 28 new papers 2 Oct–19 Oct.

What’s new

2 Oct – 19 Oct 2026

28 new papers, 2 Oct – 19 Oct; 1 this week.

  1. This week · 1
  2. AudioScorer: Learning Question-Aware Token Relevance for Large Audio Language Models
  3. Earlier · 27
  4. Probing and Mitigating Hallucinations in Speech-augmented Language Models for Automatic Speech Recognition via Small Language Models
  5. Pronouncing Hard Words: Phoneme Adaptation of a Pretrained Conversational Text to Speech Model also fits: ASR evaluation, bias and atypical speech
  6. VocalGrad: Evaluating Acoustic Perception in Audio Language Models
  7. ConvexLLM: Scaling Convex-Hull Speech Interfaces to General-Purpose Audio Intelligence
  8. TRACE: Trustworthy Reasoning from Acoustic Cues and Evidence
  9. Qwen-Audio-Jev: A Unified Interface for Audio Decisions
  10. Reference-Guided Turn-Taking Style Transfer for Full-Duplex Spoken Language Models
  11. MultiPartyTurnBench: Benchmarking Floor Control in Multi-Party Spoken Dialogue
  12. What Does LALM Post-training Learn? Rethinking Audio Understanding and Reasoning Capabilities
All 28
  1. PhonoVoxtLM: Exploring Single-Codebook Phonological Tokens for Speech Understanding and Generation
  2. Full-duplex Spoken Language Modeling with Discrete Diffusion
  3. DuplexAct: How Interactive Are Frontier Full-Duplex Spoken Language Models?
  4. D2: High-Capability Full-Duplex Spoken Language Models from Half-Duplex Backbones
  5. RefTAG: Reflection-aware Temporal Audio Grounding in Large Audio Language Models
  6. Frontend-Assisted Large Audio-Language Models for SLT2026 SmartGlasses Challenge also fits: Speaker recognition and diarization
  7. SWEAR: Section-Wise Efficient Audio Reasoning via Self-Rated Difficulty
  8. STEMMA: Song-to-Stem Multi-Audio Reasoning for Large Audio Language Models
  9. Selective Listening: Mechanism-Guided Control of Audio Influence in Large Audio-Language Models
  10. MiDashengLM-Spatial: Unifying General Audio Understanding and Spatial Awareness
  11. SmoothConv and DuplexConv: Complementary Mandarin Multi-Party Conversational Speech Corpora for Speech Interaction also fits: Multilingual and low-resource ASR
  12. BEANS-Next and ROOTS: Broadening Audio-Language Capabilities for Bioacoustics
  13. Boundary-Free Contextual Biasing: Depth-Adaptive Gating and Reading-Space Matching for Unsegmented Languages also fits: Speech recognition models and training
  14. Conversation Is a Two-Body Problem: Dyadic Evaluation of Full-Duplex Dialogue Models
  15. PERSIST: Who-What-When Memory Across Sessions for Full-Duplex Spoken Dialogue
  16. SkillFormer: Skill-Decomposed Adaptation for Audio Language Models
  17. AdaLoop: Adaptive-Depth Latent Reasoning for Audio Language Models
  18. Steering Speech-Language Models: Training-Free Task Specialization via Contrastive Activation Addition

Foundations

Most cited by papers here

  1. Qwen2-Audio Technical Report2024cited by 200 here
  2. Moshi: a speech-text foundation model for real-time dialogue2024cited by 191 here
  3. SALMONN: Towards Generic Hearing Abilities for Large Language Models2023cited by 187 here
  4. Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models2023cited by 158 here
  5. Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language Models2025cited by 136 here

Lineage

The papers this direction builds on, oldest first.

Each paper on the line cites the one before.

  1. 2021WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processingcited by 76 here
  2. 2023Neural Codec Language Models are Zero-Shot Text to Speech Synthesizerscited by 60 here
  3. 2023AudioPaLM: A Large Language Model That Can Speak and Listencited by 61 here
  4. 2023Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Modelscited by 158 here
  5. 2024AIR-Bench: Benchmarking Large Audio-Language Models via Generative Comprehensioncited by 135 here
  6. 2024Qwen2-Audio Technical Reportcited by 200 here
  7. 2024Mini-Omni: Language Models Can Hear, Talk While Thinking in Streamingcited by 81 here
  8. 2025Audio-Reasoner: Improving Reasoning Capability in Large Audio Language Modelscited by 68 here
  9. 2025MMAR: A Challenging Benchmark for Deep Reasoning in Speech, Audio, Music, and Their Mixcited by 85 here
  10. 2025Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language Modelscited by 136 here
  11. 2025Step-Audio 2 Technical Reportcited by 63 here
13 more it builds on, not in this chain
  1. 2022Generative Spoken Dialogue Language Modelingcited by 59 here
  2. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 330 here
  3. 2023AudioGPT: Understanding and Generating Speech, Music, Sound, and Talking Headcited by 79 here
  4. 2023Pengi: An Audio Language Model for Audio Taskscited by 85 here
  5. 2023SALMONN: Towards Generic Hearing Abilities for Large Language Modelscited by 187 here
  6. 2024Audio Flamingo: A Novel Audio Language Model with Few-Shot Learning and Dialogue Abilitiescited by 103 here
  7. 2024GAMA: A Large Audio-Language Model with Advanced Audio Understanding and Complex Reasoning Abilitiescited by 91 here
  8. 2024Moshi: a speech-text foundation model for real-time dialoguecited by 191 here
  9. 2024VoiceBench: Benchmarking LLM-Based Voice Assistantscited by 92 here
  10. 2024Freeze-Omni: A Smart and Low Latency Speech-to-speech Dialogue Model with Frozen LLMcited by 79 here
  11. 2024GLM-4-Voice: Towards Intelligent and Human-Like End-to-End Spoken Chatbotcited by 90 here
  12. 2025Audio Flamingo 2: An Audio-Language Model with Long-Audio Understanding and Expert Reasoning Abilitiescited by 93 here
  13. 2025MMSU: A Massive Multi-task Spoken Language Understanding and Reasoning Benchmarkcited by 64 here

The 24 most cited of 1,368.

Browse

30 of 739, ranked 6–35

  1. 6AIR-Bench: Benchmarking Large Audio-Language Models via Generative Comprehension2024135
  2. 7Audio Flamingo: A Novel Audio Language Model with Few-Shot Learning and Dialogue Abilities2024103
  3. 8Audio Flamingo 2: An Audio-Language Model with Long-Audio Understanding and Expert Reasoning Abilities202593
  4. 9VoiceBench: Benchmarking LLM-Based Voice Assistants202492
  5. 10GAMA: A Large Audio-Language Model with Advanced Audio Understanding and Complex Reasoning Abilities202491
  6. 11GLM-4-Voice: Towards Intelligent and Human-Like End-to-End Spoken Chatbot202490
  7. 12MMAR: A Challenging Benchmark for Deep Reasoning in Speech, Audio, Music, and Their Mix202585
  8. 13Pengi: An Audio Language Model for Audio Tasks202385
  9. 14Mini-Omni: Language Models Can Hear, Talk While Thinking in Streaming202481
  10. 15Freeze-Omni: A Smart and Low Latency Speech-to-speech Dialogue Model with Frozen LLM202479
All 30
  1. 16AudioGPT: Understanding and Generating Speech, Music, Sound, and Talking Head202379
  2. 17Audio-Reasoner: Improving Reasoning Capability in Large Audio Language Models202568
  3. 18MMSU: A Massive Multi-task Spoken Language Understanding and Reasoning Benchmark202564
  4. 19Step-Audio 2 Technical Report202563
  5. 20AudioPaLM: A Large Language Model That Can Speak and Listen202361
  6. 21Generative Spoken Dialogue Language Modeling202259
  7. 22Full-Duplex-Bench: A Benchmark to Evaluate Full-duplex Spoken Dialogue Models on Turn-taking Capabilities202557
  8. 23Beyond Turn-Based Interfaces: Synchronous LLMs as Full-Duplex Dialogue Agents202448
  9. 24WavChat: A Survey of Spoken Dialogue Models202444
  10. 25Spirit LM: Interleaved Spoken and Written Language Model202443
  11. 26SD-Eval: A Benchmark Dataset for Spoken Dialogue Understanding Beyond Words202443
  12. 27Spoken Question Answering and Speech Continuation Using Spectrogram-Powered LLM202341
  13. 28Audio-CoT: Exploring Chain-of-Thought Reasoning in Large Audio Language Model202540
  14. 29MiMo-Audio: Audio Language Models are Few-Shot Learners202540
  15. 30DeSTA2.5-Audio: Toward General-Purpose Large Audio Language Model with Self-Generated Cross-Modal Alignment202539
  16. 31Language Model Can Listen While Speaking202439
  17. 32PersonaPlex: Voice and Role Control for Full Duplex Conversational Speech Models202637
  18. 33OmniFlatten: An End-to-end GPT Model for Seamless Voice Conversation202437
  19. 34Recent Advances in Speech Language Models: A Survey202436
  20. 35Full-Duplex-Bench v1.5: Evaluating Overlap Handling for Full-Duplex Speech Models202534