Microsoft open-sources VibeVoice-ASR-Streaming with real-time diarization in 10 languages
realmrfakename · x · 2026-09-03
Microsoft released VibeVoice-ASR-Streaming, an open-source streaming ASR model that continuously transcribes "who said what" as speech arrives, combining transcription and diarization in one model with hotword support across 10 languages. A 7B checkpoint is on Hugging Face; the GitHub repo has 53k+ stars. Related: VibeVoice-ASR-BitNet compresses the model from 4.62GB to 1.58GB via heterogeneous quantization (I8S + I2S) for real-time CPU-only inference (RTF < 1, 3+ threads), and VibeVoice ASR is already integrated into Azure AI Foundry Labs and Transformers. Third parties say it's competitive with closed-source models.
More from Models
- Users press OpenAI on long-delayed Adult Mode a year after it was promised — ___Patrice___ · 2026-10-18
- Polymarket opens betting market on who releases the next Gemini Argon model — Polymarket · 2026-10-18
- Grok matter-of-factly lists which groups user was racist against, exempting Africans and Latin Americans — teortaxesTex · 2026-10-18
- A quick test for benchmaxxed demos: swap in Mirror's Edge and see if it generalizes — teortaxesTex · 2026-10-18
- Did OpenAI distill DeepSeek's gains? Benchmark shows OpenAI's Luna scoring just 38 — teortaxesTex · 2026-10-18
- Gemini 4 reportedly near Opus 5.5 on complex tasks, surprisingly strong at security — haider1 · 2026-10-18