Confucius4-R2T2: Stable-Prefix ASR Model for Real-Time Transcription

News

Confucius4-R2T2 introduces stable-prefix streaming ASR with high accuracy and low latency, enabling real-time transcription without output revisions.

Researchers have unveiled Confucius4-R2T2, a stable-prefix streaming automatic speech recognition (ASR) model designed for real-time transcription. Published on November 10, 2026, on arXiv and highlighted on HuggingFace Daily Papers, the model introduces a significant advancement in streaming ASR by ensuring that once text is emitted, it remains immutable, all while maintaining high accuracy and low latency.

Technical Innovations and Training Methodology

Confucius4-R2T2, short for Real Real-Time Transcription, adapts the Qwen3-ASR-1.7B model through post-training on 1,500 hours of Chinese and English audio. The training process employs the Longest Stable Prefix (LSP) learning method, a novel concept introduced by the researchers. LSP ensures that the model’s output is monotonic, consistent with full-audio transcription, and maximally stable. This approach allows the model to convert from a full-context ASR to a stable-prefix streaming ASR without architectural modifications.

The training process freezes the pre-trained audio encoder and projector, updating only the language decoder through two-stage fine-tuning using Supervised Fine-Tuning (SFT) and Gradient Projection Optimization (GRPO). Additionally, the model incorporates an explicit ‘Wait’ action to determine when more audio is needed before committing to a transcript.

Performance Metrics

Confucius4-R2T2 delivers substantial improvements in streaming recognition accuracy. At a latency of 160 milliseconds with one provisional token, the model achieves a recognition error rate of 5.76%, significantly outperforming the base model’s 21.01% error rate under the same conditions. This positions R2T2 as a leading open-source solution for low-latency, high-accuracy transcription.

Applications and Real-World Impact

The model’s stable-prefix output is particularly beneficial for applications requiring immediate and reliable text processing, such as live captioning, simultaneous speech translation, and downstream natural language processing pipelines. By eliminating disruptive text revisions and visual flicker, R2T2 enhances the user experience in real-time environments.

Open-Source Availability

To encourage further research and adoption, the developers have open-sourced the model weights and code. The resources are available on HuggingFace (netease-youdao/Confucius4-R2T2) and GitHub (netease-youdao/Confucius4-R2T2).

Confucius4-R2T2 represents a practical pathway from full-context ASR to stable streaming, offering a robust solution for real-time transcription needs across multiple languages and use cases.