AliQwen-Audio-3.0 Series Speech Models Launched on Tongyi AI Platform, Winning a Grand Slam in International Evaluations

Recently, Alibaba officially announced the full launch of the Qwen-Audio-3.0 series speech models on the Qwen AI platform. The series includes the large-scale speech recognition model Qwen-Audio-3.0-ASR, the large-scale speech synthesis model Qwen-Audio-3.0-TTS, and the real-time speech interaction dialogue model Qwen-Audio-3.0-Realtime.

According to the official introduction, the series models performed excellently in the voice rankings of the authoritative AI evaluation platform Artificial Analysis in July this year, winning first place in three core categories: speech recognition (ASR), real-time interaction (RealTime), and speech synthesis (TTS), achieving a "grand slam".

In terms of capabilities, Qwen-Audio-3.0-ASR supports identification in complex contexts and professional fields; Qwen-Audio-3.0-TTS supports multiple languages and dialects, and can accurately control emotions, tone, and rhythm; Qwen-Audio-3.0-Realtime realizes end-to-end speech understanding and dialogue, supporting interrupting and asking follow-up questions at any time, as well as tool calls, and has been applied to products such as Qwen App, Qwen Office, and Qoder.

Additionally, the Qwen AI platform has recently updated multiple main models, including Qwen3.8-Max, Qwen-image3.0pro, and Wan3.0, forming a comprehensive multi-modal model matrix covering text, code, speech, images, and video. The platform also launched a Token Plan subscription service with limited-time offers, starting at 39 yuan/month for the personal plan, and up to a 7.6 discount for the team plan, aiming to lower the access threshold for developers and create an AI productivity platform "born for Agents."

Qwen Launches a Major Upgrade: Real-Time Speech Recognition Model Fun-ASR-Realtime Officially Released

Qwen introduces the real-time speech recognition model Fun-ASR-Realtime, reducing the first-word latency to the millisecond level and achieving a smooth "speak and feedback immediately" interaction. Its recognition accuracy is close to that of offline models, achieving high precision while breaking through the real-time performance bottleneck, marking a new height in voice interaction experience.

iFlytek Launches AI Hardware and Software Integrated Solution: Accurate Recognition Even in 90 Decibel Noise

iFlytek launched its AI hardware and software integrated solution at the 2025 1024 Developer Festival. By deeply integrating algorithms and hardware, it solves recognition challenges in complex environments such as high noise and far-field conditions, improving the accuracy of voice and visual intelligence, marking a significant breakthrough in this field.

Mianbi Intelligence Launches VoxCPM: A Next-Generation High-Fidelity Speech Generation Model

In the context of rapid development in speech synthesis technology, Mianbi Intelligence and the Human-Computer Speech Interaction Laboratory at the Shenzhen International Graduate School, Tsinghua University (THUHCSI) recently jointly released a new speech generation model - VoxCPM. This model, with a parameter size of 0.5B, is dedicated to providing users with high-quality and natural speech synthesis experiences. The release of VoxCPM marks another milestone in the field of high-fidelity speech generation. The model has achieved industry-leading levels in key indicators such as naturalness, voice similarity, and prosodic expressiveness.

Alibaba Launches Revolutionary Speech Recognition Model FunAudio-ASR with Remarkable Noise Reduction

Recently, Tongyi Lab of Alibaba officially released its latest end-to-end speech recognition large model - FunAudio-ASR. The biggest highlight of this model is its innovative "Context Module," which significantly improves the accuracy of speech recognition in high-noise environments. The hallucination rate has been reduced from 78.5% to 10.7%, a decrease of nearly 70%. This technological breakthrough has set a new benchmark for the speech recognition industry, especially suitable for noisy environments such as meetings and public places. FunAudio-AS

OpenAI Evals Adds Native Audio Input and Evaluation Features

Recently, the OpenAI Evals tool received a significant and exciting update, adding native audio input and evaluation features. This innovation means that developers can now evaluate speech recognition and generation models directly using audio files, without going through the cumbersome process of text transcription. This change greatly simplifies the evaluation process, making the development of audio applications more efficient. In previous evaluation processes, developers often needed to first convert audio content into text, which was time-consuming and labor-intensive, and could also affect