# video-SALMONN 2: Caption-Enhanced Audio-Visual Large Language Models

arXiv: https://arxiv.org/abs/2506.15220 · HTML: https://mlsanity.com/paper/2506.15220/

## Key facts

- Cited by 47 papers in the map (data through 14 Oct 2026).
- Builds on 51 of its 62 references that are in the map.
- First posted on arXiv 18 Jun 2025.

## Builds on (most cited first, 30 of 51)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [Finetuned Language Models Are Zero-Shot Learners](https://mlsanity.com/paper/2109.01652/) (2021)
- [The Curious Case of Neural Text Degeneration](https://mlsanity.com/paper/1904.09751/) (2019)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks](https://mlsanity.com/paper/2312.14238/) (2023)
- [Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond](https://mlsanity.com/paper/2308.12966/) (2023)
- [Video-LLaMA: An Instruction-tuned Audio-Visual Language Model for Video Understanding](https://mlsanity.com/paper/2306.02858/) (2023)
- [Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis](https://mlsanity.com/paper/2405.21075/) (2024)
- [Video-LLaVA: Learning United Visual Representation by Alignment Before Projection](https://mlsanity.com/paper/2311.10122/) (2023)
- [VideoLLaMA 2: Advancing Spatial-Temporal Modeling and Audio Understanding in Video-LLMs](https://mlsanity.com/paper/2406.07476/) (2024)
- [Qwen2.5-Omni Technical Report](https://mlsanity.com/paper/2503.20215/) (2025)
- [VILA: On Pre-training for Visual Language Models](https://mlsanity.com/paper/2312.07533/) (2023)
- [Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models](https://mlsanity.com/paper/2311.07919/) (2023)
- [LLaVA-Video: Video Instruction Tuning With Synthetic Data](https://mlsanity.com/paper/2410.02713/) (2024)
- [Qwen2-Audio Technical Report](https://mlsanity.com/paper/2407.10759/) (2024)
- [VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understanding](https://mlsanity.com/paper/2501.13106/) (2025)
- [SALMONN: Towards Generic Hearing Abilities for Large Language Models](https://mlsanity.com/paper/2310.13289/) (2023)
- [LVBench: An Extreme Long Video Understanding Benchmark](https://mlsanity.com/paper/2406.08035/) (2024)
- [ShareGPT4Video: Improving Video Understanding and Generation with Better Captions](https://mlsanity.com/paper/2406.04325/) (2024)

## Built on by (most cited first, 30 of 47)

- [JointAVBench: A Benchmark for Joint Audio-Visual Reasoning Evaluation](https://mlsanity.com/paper/2512.12772/) (2025)
- [OmniVideo-R1: Reinforcing Audio-visual Reasoning with Query Intention and Modality Attention](https://mlsanity.com/paper/2602.05847/) (2026)
- [OmniPro: A Comprehensive Benchmark for Omni-Proactive Streaming Video Understanding](https://mlsanity.com/paper/2605.18577/) (2026)
- [Script-a-Video: Deep Structured Audio-visual Captions via Factorized Streams and Relational Grounding](https://mlsanity.com/paper/2604.11244/) (2026)
- [OmniVideo-100K: A Dataset for Audio-Visual Reasoning through Structured Scripts and Evidence Chains](https://mlsanity.com/paper/2606.14702/) (2026)
- [FutureOmni: Evaluating Future Forecasting from Omni-Modal Context for Multimodal LLMs](https://mlsanity.com/paper/2601.13836/) (2026)
- [On the Nature of Attention Sink that Shapes Decoding Strategy in Omni-LLMs](https://mlsanity.com/paper/2603.14337/) (2026)
- [OmniScript: Towards Audio-Visual Script Generation for Long-Form Cinematic Video](https://mlsanity.com/paper/2604.11102/) (2026)
- [Don't Let the Video Speak: Audio-Contrastive Preference Optimization for Audio-Visual Language Models](https://mlsanity.com/paper/2604.14129/) (2026)
- [A Benchmark for Omni-Modal Reasoning in Long Videos](https://mlsanity.com/paper/2512.16978/) (2025)
- [AVERE: Improving Audiovisual Emotion Reasoning with Preference Optimization](https://mlsanity.com/paper/2602.07054/) (2026)
- [JAEGER: Joint 3D Audio-Visual Grounding and Reasoning in Simulated Physical Environments](https://mlsanity.com/paper/2602.18527/) (2026)
- [Watch, Remember, Reason: Human-View Video Understanding with MLLMs](https://mlsanity.com/paper/2606.07433/) (2026)
- [Realtime-Venus: A full-duplex interaction system with asynchronous delegation](https://mlsanity.com/paper/2609.13814/) (2026)
- [Keep What Audio Cannot Say: Context-Preserving Token Pruning for Omni-LLMs](https://mlsanity.com/paper/2605.11605/) (2026)
- [Stage-adaptive Token Selection for Efficient Omni-modal LLMs](https://mlsanity.com/paper/2605.20035/) (2026)
- [AVOC: Enhancing Hour-Level Audio-Video Understanding in Omni-Modal LLMs via Retrieval-Inspired Token Compression](https://mlsanity.com/paper/2606.24286/) (2026)
- [Temporal and Cross-Modal Alignment for Enhanced Audiovisual Video Captioning](https://mlsanity.com/paper/2607.01667/) (2026)
- [ReBind: Multi-Reference Video Editing via Structured Instructions with Explicit Reference Relationships](https://mlsanity.com/paper/2607.14681/) (2026)
- [OmniPack: Unified Token Compression for Efficient Omni-modal Large Language Models](https://mlsanity.com/paper/2608.03812/) (2026)
- [PersonalHomeBench: Evaluating Agents in Personalized Smart Homes](https://mlsanity.com/paper/2604.16813/) (2026)
- [Separate First, Fuse Later: Mitigating Cross-Modal Interference in Audio-Visual LLMs Reasoning with Modality-Specific Chain-of-Thought](https://mlsanity.com/paper/2605.09906/) (2026)
- [SVHalluc: Benchmarking Speech-Vision Hallucination in Audio-Visual Large Language Models](https://mlsanity.com/paper/2606.02642/) (2026)
- [OmniCap-IF: Benchmarking and Improving Instruction Following Abilities for Omni-Video Captioning](https://mlsanity.com/paper/2606.08572/) (2026)
- [AVSCap: Orchestrating Audio-Visual Synergy for Omni-modal Video Captioning](https://mlsanity.com/paper/2607.12820/) (2026)
- [Audio-Visual Flamingo: Open Audio-Visual Intelligence for Long and Complex Videos](https://mlsanity.com/paper/2607.16107/) (2026)
- [TraceAV-Bench: Benchmarking Multi-Hop Trajectory Reasoning over Long Audio-Visual Videos](https://mlsanity.com/paper/2605.07593/) (2026)
- [VCap: Hypergeometric Rewards for Weak-to-Strong Visual Captioning](https://mlsanity.com/paper/2605.28023/) (2026)
- [From Senses to Decisions: The Information Flow of Auditory and Visual Perception in Multimodal LLMs](https://mlsanity.com/paper/2606.10147/) (2026)
- [OmniScope: Modality-Decoupled Token Compression for Omnimodal Large Language Models](https://mlsanity.com/paper/2607.23193/) (2026)

## Abstract (arXiv)

We present video-SALMONN 2, a family of audio-visual large language models that set new state-of-the-art (SOTA) results in video description and question answering (QA). Our core contribution is multi-round direct preference optimisation (MrDPO), paired with a caption-quality objective that jointly rewards completeness and factual accuracy. Unlike standard DPO with a fixed reference policy, MrDPO periodically refreshes the reference by bootstrapping from a newly re-initialised lightweight adapter trained on the latest preferences, avoiding reference staleness and enabling continual improvement. This strategy produces captions that are consistently more detailed and accurate than those from proprietary systems such as GPT-4o and Gemini-1.5 Pro. We further distil these gains by using our model to generate a high-quality video-caption corpus for supervised fine-tuning of new models, transferring benefits beyond captioning to strong performance on complex video-QA tasks. Across widely used audio-visual and visual-only understanding benchmarks (including Video-MME, WorldSense, AVUT, Video-Holmes, DailyOmni, MLVU, and LVBench), our 3B and 7B models achieve SOTA results at comparable scales, while the 72B model surpasses all other open-source systems. Our source code, models, and data are released at hrefhttps://github.com/bytedance/video-SALMONN-2https://github.com/bytedance/video-SALMONN-2.