Microsoft Research Unveils VibeVoice for Long-Form Speech Synthesis
Microsoft’s VibeVoice is an open-source text-to-speech model that generates podcast-length audio with up to four...
Microsoft’s VibeVoice is an open-source text-to-speech model that generates podcast-length audio with up to four...
Free, open-source transcription app Whispering lets users keep recordings on-device or connect directly to providers...
Cohere releases Command A Translate, a new large language model built specifically for AI translation,...
New study finds that prompt engineering has limits in AI translation. If a large language...
The WMT25 preliminary results are out: Gemini-2.5-Pro and GPT-4.1 lead AI translation. Commercial engines remain...
China’s ByteDance is building an end-to-end language AI stack — spanning AI text and live...
Researchers from Sapienza University, ETH Zurich, and Cohere introduce “translation difficulty estimation,” a method designed...
Welocalize and Duke University release a new dataset built to evaluate the post-editing capabilities of...
Researchers from German university hospitals find GPT-4o can provide reliable medical report translations as templates...
At IWSLT 2025, researchers and industry share ways to boost speed, quality, subtitles, model size,...
OpenAI’s new GPT-OSS models run locally, offline, and on everything from laptops to enterprise servers,...