# Speed by Simplicity: A Single-Stream Architecture for Fast Audio-Video Generative Foundation Model

arXiv: https://arxiv.org/abs/2603.21986 · HTML: https://mlsanity.com/paper/2603.21986/

## Key facts

- Cited by 25 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 18 references that are in the map.
- First posted on arXiv 23 Mar 2026.

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

- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [HunyuanVideo: A Systematic Framework For Large Video Generative Models](https://mlsanity.com/paper/2412.03603/) (2024)
- [Improved Distribution Matching Distillation for Fast Image Synthesis](https://mlsanity.com/paper/2405.14867/) (2024)
- [Gated Attention for Large Language Models: Non-linearity, Sparsity, and Attention-Sink-Free](https://mlsanity.com/paper/2505.06708/) (2025)
- [LTX-2: Efficient Joint Audio-Visual Foundation Model](https://mlsanity.com/paper/2601.03233/) (2026)
- [Ovi: Twin Backbone Cross-Modal Fusion for Audio-Video Generation](https://mlsanity.com/paper/2510.01284/) (2025)
- [UniVerse-1: Unified Audio-Video Generation via Stitching of Experts](https://mlsanity.com/paper/2509.06155/) (2025)
- [Is Noise Conditioning Necessary for Denoising Generative Models?](https://mlsanity.com/paper/2502.13129/) (2025)
- [MOVA: Towards Scalable and Synchronized Video-Audio Generation](https://mlsanity.com/paper/2602.08794/) (2026)
- [VideoScore2: Think before You Score in Generative Video Evaluation](https://mlsanity.com/paper/2509.22799/) (2025)
- [Turbo-VAED: Fast and Stable Transfer of Video-VAEs to Mobile Devices](https://mlsanity.com/paper/2508.09136/) (2025)

## Built on by (most cited first, 25 of 25)

- [Hallo-Live: Real-Time Streaming Joint Audio-Video Avatar Generation with Asynchronous Dual-Stream and Human-Centric Preference Distillation](https://mlsanity.com/paper/2604.23632/) (2026)
- [MSAVBench: Towards Comprehensive and Reliable Evaluation of Multi-Shot Audio-Video Generation](https://mlsanity.com/paper/2605.20183/) (2026)
- [LongAV-Compass: Towards Unified Evaluation of Minute-Scale Audio-Visual Generation Across T2AV, I2AV, and V2AV](https://mlsanity.com/paper/2605.26244/) (2026)
- [Script-a-Video: Deep Structured Audio-visual Captions via Factorized Streams and Relational Grounding](https://mlsanity.com/paper/2604.11244/) (2026)
- [Do Joint Audio-Video Generation Models Understand Physics?](https://mlsanity.com/paper/2605.07061/) (2026)
- [Native Audio-Visual Alignment for Generation](https://mlsanity.com/paper/2605.30073/) (2026)
- [VA-Judger: Reward Modeling from Human Preference Feedback for Joint Video-Audio Generation](https://mlsanity.com/paper/2608.18607/) (2026)
- [Artifact-Bench: Evaluating MLLMs on Detecting and Assessing the Artifacts of AI-Generated Videos](https://mlsanity.com/paper/2605.18984/) (2026)
- [StreamChar: Long-Horizon Streaming Character Audio-Video Generation with Decoupled Orchestration](https://mlsanity.com/paper/2605.25659/) (2026)
- [Ultra Flash: Scaling Real-Time Streaming Video Generation to High Resolutions](https://mlsanity.com/paper/2606.09150/) (2026)
- [Salt++: Context-Aligned Post-Training for Few-Step Streaming Multimodal Generation](https://mlsanity.com/paper/2609.36995/) (2026)
- [KeyFrame-Compass: Towards Comprehensive Evaluation of Keyframe-Conditioned Video Generation](https://mlsanity.com/paper/2607.14202/) (2026)
- [Ripple: Real-Time Streaming Audio-Video Generation With Cross-Modal Recurrent Memory](https://mlsanity.com/paper/2607.26818/) (2026)
- [Avatar-Forever: Decoupled Parallel Training for High-Quality Real-Time Infinite Avatars](https://mlsanity.com/paper/2608.12107/) (2026)
- [Omni-LiveAvatar: Minute-Level Real-Time Streaming Joint Audio-Video Avatar Generation](https://mlsanity.com/paper/2608.13602/) (2026)
- [Beyond Coherence: Benchmarking Professional Editing-Technique Execution in Multi-Shot Audio-Video Generation](https://mlsanity.com/paper/2609.08275/) (2026)
- [OmniMate: Open-Ended Real-Time Streaming Audio-Visual Generation for Interactive Avatars](https://mlsanity.com/paper/2607.23023/) (2026)
- [Vorch-Omni: Multi-Task Orchestration of Sight and Sound](https://mlsanity.com/paper/2608.05803/) (2026)
- [TurboT2VA: Fast Large-Scale Text-to-Video-Audio Generation via Score-Regularized Consistency Distillation](https://mlsanity.com/paper/2608.24674/) (2026)
- [Multi2AV-Safety: Benchmarking Safety in Multimodal-to-Audio-Video Generation](https://mlsanity.com/paper/2608.26535/) (2026)
- [DreamX-Creator: Democratizing Native Audio-Video Generation at 2K Resolution](https://mlsanity.com/paper/2608.31106/) (2026)
- [TimeSteer: Inference-Time Speech Scheduling in Joint Audio-Visual Diffusion Models](https://mlsanity.com/paper/2609.01277/) (2026)
- [Vorch-Human: Unified Multi-Task Human-Centric Generation via Long-Horizon Continuation](https://mlsanity.com/paper/2609.26117/) (2026)
- [Prism: Dynamic Sparse Attention for Native 2K Joint Video-Audio Generation Model Training](https://mlsanity.com/paper/2610.05416/) (2026)
- [SyncReward: Human-Aligned Synchronization Reward for Audio-Visual Generation](https://mlsanity.com/paper/2610.16338/) (2026)

## Abstract (arXiv)

We present daVinci-MagiHuman, an open-source audio-video generative foundation model for human-centric generation. daVinci-MagiHuman jointly generates synchronized video and audio using a single-stream Transformer that processes text, video, and audio within a unified token sequence via self-attention only. This single-stream design avoids the complexity of multi-stream or cross-attention architectures while remaining easy to optimize with standard training and inference infrastructure. The model is particularly strong in human-centric scenarios, producing expressive facial performance, natural speech-expression coordination, realistic body motion, and precise audio-video synchronization. It supports multilingual spoken generation across Chinese (Mandarin and Cantonese), English, Japanese, Korean, German, and French. For efficient inference, we combine the single-stream backbone with model distillation, latent-space super-resolution, and a Turbo VAE decoder, enabling generation of a 5-second 256p video in 2 seconds on a single H100 GPU. In automatic evaluation, daVinci-MagiHuman achieves the highest visual quality and text alignment among leading open models, along with the lowest word error rate (14.60%) for speech intelligibility. In pairwise human evaluation, it achieves win rates of 80.0% against Ovi 1.1 and 60.9% against LTX 2.3 over 2000 comparisons. We open-source the complete model stack, including the base model, the distilled model, the super-resolution model, and the inference codebase.