Ant Group inclusionAI · Hugging Face new modelsinclusionAI has released the Duo 1.2 causal-base model, a full-duplex multimodal world model for real-time audio-video interaction, featuring improved generation quality, and model weights can be downloaded from Hugging Face.
Ant Group inclusionAI · Hugging Face new modelsinclusionAI releases Duo-1.0-distill, a full-duplex multimodal world model for real-time audio-video interaction, including Realization weights with four-step distillation, which needs to be used in conjunction with the causal-base suite.
AWS Machine Learning BlogAmazon SageMaker AI Two-Way Stream introduces a pay-as-you-go pricing model, enabling model providers to report actual usage of real-time streaming inference through the AWS Marketplace. Deepgram becomes the first partner to bring voice-to-text and text-to-speech pay-as-you-go to the AWS Marketplace.
Google Developers BlogGoogle AI Edge Gallery has launched on-device automatic speech recognition (ASR) and text-to-speech (TTS) capabilities, enabling fully offline operation of multilingual ASR and natural TTS models on phones and wearable devices.
Claude · YouTube5858Claude Fable 5.2 designs a voice assistant center from a prompt word, including circuit board layout, CAD shell modeling, and firmware writing. Diode Computers manufactured the prototype board according to its design documents. This is one of the three devices designed by Claude in the project. Anthropic and Diode have a business relationship.
Cursor ChangelogFeatured6060Cursor launches Voice Chat, allowing users to have real-time voice calls with agents, dictate modification requests, ask what the agent is doing, and receive progress updates.
Why it matters: Officially, it explains which platforms the voice call is available on and how to divide tasks with coding agents, allowing readers to judge which collaboration scenarios it suits.
Hugging Face BlogHugging Face Open ASR leaderboard adds Liva AI private datasets for Egyptian English and Nigerian English, containing 1.34 hours and 1.55 hours of dialogue speech, to enhance the representativeness and reliability of ASR evaluation. The leaderboard now supports more detailed accent classification based on the Kachru three-circle model, covering groups such as US and non-US, inner circle and outer circle/extended circle.
Hugging Face BlogFeatured7272Phonon releases a small TTS model that can run on local devices, without a server and network, offering two versions of about 65 MB and 75 MB, covering English, French, German, Spanish, and Portuguese.
Why it matters: The original text outlines the specifications, open-source license, and leaderboard performance of on-device TTS models, allowing judgment on the usability boundary of small models in local voice scenarios.
AWS Machine Learning Blog3838Cartesia uses state space models (SSM) designed for real-time speech AI and AWS infrastructure to achieve sub-100ms text-to-speech and speech-to-text, supporting voice cloning for 44 languages, suitable for scenarios such as virtual assistants and customer service centers. The Sonic model generates the first audio byte in about 90ms, and the Ink model has native turn detection and noise robustness. The API interface supports real-time streaming output.
Hugging Face Blog4040MonsterHouse is an interactive projection built with NVIDIA Nemotron Voice Agent, NVIDIA Nemotron ASR, NVIDIA Nemotron 3.5 Bolt E4B 4GB, Magpie TTS Zeroshot, and Audio2Face-3D, which can communicate with visitors through voice and imitate the voice of a grumpy old pirate.
Codex releases4545Codex 0.166.0 released, adding keyboard driver transcription selection (supporting word, line, and page operations), integrated search and configurable shortcuts, as well as consistent Markdown rendering in asynchronous questions, options, answers, and transcription responses. Also supports GPT-6.1 Terra and adds codex queue --resume feature to resume unloaded threads.
Hugging Face Blog5050Hugging Face releases open-source speech recognizer AnySpeechRecognizer and decision model AnyDecisionModel, supporting NVIDIA Parakeet and Alibaba Cloud Qwen3-ASR engines, and can run locally on Apple platforms.
Databricks BlogDatabricks releases voice intelligence agents, which adjust responses in conversations based on real-time contextual decisions, enabling language switching, transferring to a human, or providing personalized services. These agents need to process customer requests within 200 milliseconds and rely on the context, ontology, and governance capabilities provided by Databricks. They integrate customer, order, product, and policy data, and follow the access rules of the enterprise data catalog.
Hugging Face BlogFeatured7878NVIDIA has released the open-source weight model Nemotron 3.5 Transcribe, integrating transcription, speaker separation, translation, and audio Q&A into a single checkpoint, supporting transcription in 40 languages, translation between 31 languages and English, and speaker annotation in 13 languages, capable of processing the longest 1-hour recordings and annotating up to 8 speakers.
Why it matters: The original text provides a unified weighting model for multilingual transcription, speaker separation, and translation, along with an open data pipeline and fine-tuning recipes, which can be used to determine the adaptation path for voice scenarios.
PyTorch Blog3434Nemotron 3.5 achieved the highest streaming speech recognition accuracy on Google Pixel 11 using ExecuTorch and SME2, reaching up to 11.7% WER (with 1,120 ms audio blocks and QC8W precision), which is 3.5x faster than FP32 but with increased latency; shorter audio blocks (e.g., 80 ms) can reduce latency but accuracy drops to 15.2% WER.
Ollama BlogFeatured6262Ollama adds text-to-speech capability, the first integrated speech model is the small TTS model family Phonon from Gradium, currently supporting macOS, Windows, and Linux with support for Linux to be released soon.
Why it matters: The original text provides throughput comparison, accuracy, and OpenAI-compatible interface after Ollama integrates speech models, which can be used to judge the local TTS implementation method.
Amazon ScienceAudible's technology team has developed a new controllable speech synthesis technology, by combining large language models (LLMs) and three innovative technologies to enhance the experience of audiobooks: 1) Develop a data synthesis pipeline to generate speech data with voice feature descriptions, train the model to understand the descriptions and generate matching audio; 2) Create an automatic system to analyze book characters and identify dialogues, with an accuracy of 97%, to provide stage directions for the model; 3) Improve the dialogue discrimination by 3.7 points in MUSHRA tests, and significantly improve the sound experience in English and Spanish tests by 80%.
AWS Machine Learning Blog4848Amazon Polly Generative engine upgrade provides more natural and accurate speech output, adding 14 new voices (across 9 regions), for the first time making Hindi and Finnish generally available. The upgrade is plug-and-play, with unchanged API calls and voice names, so existing applications do not need to be modified. Supports emotion and prosody expression based on text context, with new languages being Hindi (Arjun/Kiara) and Finnish (Suvi), totaling 57 generated voices (11 languages/25 regions), with prices unchanged.
AWS Machine Learning BlogTencent Hunyuan · Hugging Face new modelsFeatured7272Tencent Mixuan has released and open-sourced the real-time simultaneous translation model Hy-SimulMT-30B-A3B, supporting bidirectional Chinese-English voice-to-voice translation, capable of continuously outputting translated text and speech during speech, and supporting real-time voice cloning.
Why it matters: The official provides a comparison of latency and quality for real-time simultaneous translation, and opens up multiple quantitative versions and inference deployment scripts, which can be used to evaluate the deployment cost.
Mistral AIMistral AI has launched Voxtral 3, which includes an advanced voice model capable of real-time multi-step conversations in over 50 languages, along with a voice agent SDK. Its voice-to-text model, Voxtral Small Transcribe 3, achieved a Forced WER of 0.9% on the FLEURS 45 benchmark.
Google Developers BlogUsing ADK and Gemini 3.8, non-blocking voice interaction is implemented through streaming tools and asynchronous functions, reducing the first response time of the sales report workflow from 14.4 seconds to 1.2 seconds, supporting zero silence, progress updates, and parallel conversations.
Ant Group inclusionAI · Hugging Face new modelsinclusionAI/Ming-3.0-asr-tiny combines Ling-tiny-v3 language backbone and audio encoder, supporting single-round Chinese and English speech transcription and recognition of 15 dialects, improves professional term recognition rate through hot words or dialogue context, model parameter size not mentioned, evaluated in noisy speech and 5 application areas, provides PyTorch/Transformers inference examples and MIT license.
Apple Machine Learning Research2727Whisper-small achieved an average F1 score of 0.798 across 14 language pairs in the CS-FLEURS benchmark test, surpassing CNN's 0.776 and HuBERT's linear baseline's 0.751. The method detects language switching directly from speech by identifying the language-selective hidden state dimension in pre-trained speech encoders, and only requires a small amount of code to convert the dataset to fine-tune decision rules.
OpenAI News5555Higgsfield launches GPT-6 Astra, GPT-Image-2.5, and GPT-Live, supporting creators in making longer videos, achieving precise visual effects, and providing voice movie production paths.
Microsoft AI NewsMAI-Transcribe-2-Streaming provides low-latency real-time transcription for 60 languages and supports automatic language detection, achieving the top accuracy in both final and partial transcription on Artificial Analysis.
Google DeepMindGoogle DeepMind releases Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, supporting the creation of custom character voices from scratch and direct control of scene dialogue.
Google DeepMindGoogle DeepMind has launched the Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking large language models, with the former designed specifically for scale and cost-effectiveness.
Microsoft AI NewsMAI-Transcribe-2 achieved an average Word-Error-Rate of 5.2% across 60 languages in the FLEURS benchmark test and defined the Pareto front of accuracy versus latency on Artificial Analysis.
Microsoft AI NewsMAI's audio model Ode Poetry recommends and plays real human reading recordings of poems for users worldwide, providing a digital poetry experience. The system uses MAI-Transcribe to understand user language in real-time, achieving the lowest word error rate of 4.86% in the FLEURS benchmark test; MAI-Voice can faithfully reproduce the unique voice of William Sieghart. The system combines William's literary analysis experience, recommends poems through custom logic and tool calls, and iteratively optimizes for a safe, high-quality user experience.
Microsoft AI NewsMAI-Transcribe-1 achieved the lowest Word Error Rate for 25 languages in the FLEURS evaluation, outperforming competitors such as Scribe v2 and Whisper-large-V3, while maintaining high accuracy in 25 languages. The model offers a batch transcription speed 2.5 times faster than Microsoft's existing Azure Fast, priced at $0.36 per hour, and is now available in Microsoft Foundry for open preview.
MiniMax NewsMiniMax Speech 2.8 released, enhancing voice realism through native sound tags, high-fidelity cloning, and studio-grade clarity. It can precisely clone user voice features within 10 seconds, supporting background noise removal and cross-language pronunciation optimization.
