Model reference · open weights
VibeVoice-ASR-BitNet is an open-weight audio or speech model from microsoft, listed in the AxForge catalogue. AxForge can bring it up on EU-owned hardware for you on request — with the licence handled where one is required.
About
VibeVoice-ASR-BitNet [](https://github.com/microsoft/VibeASR.cpp) [](https://arxiv.org/abs/2607.21075) [](https://opensource.org/licenses/MIT) VibeVoice-ASR-BitNet is a compressed variant of VibeVoice-ASR optimized for real-time inference on edge CPUs — no GPU required. Through heterogeneous quantization, the model is compressed from 4.62 GB to 1.58 GB while achieving 1.6–2.3× faster inference than Whisper.cpp with real-time capability (RTF < 1) on as few as 3 CPU threads. ➡️ Code: microsoft/VibeASR.cpp<br ➡️ Report: VibeVoice-ASR-BitNet Technical Report<br ➡️ Base Model: microsoft/VibeVoice-ASR<br 🔥 Key Features - ⚡ Real-time on CPU — RTF < 1 with 3+ threads on commodity x86 (AVX2) and ARM (NEON) hardware - 📦 Compact — 1.58 GB total (2.9× compression from FP16), fits in edge device memory - 🌍 Multilingual — English, Chinese, French, Italian, Korean, Portuguese, Vietnamese, and more - 🔧 Custom SIMD Kernels — Fused operators within the ggml framework for both ARM and x86 platforms Quantization Strategy Evaluation Inference Speed Benchmarked on AMD EPYC 7V13 (AVX2+FMA) with 20s audio. Bold = RTF < 1 (real-time). Accuracy (WER%) Model Files License This project is licensed under the MIT License. Contact This project was conducted by members of Microsoft Research. If you have suggestions, questions, or observe unexpected behavior, please contact us at VibeVoice@microsoft.com.
Summarised from the published model card. Read the full card on the HuggingFace links below.
Specifications
| Maker | microsoft |
|---|---|
| Type | Audio & music |
| Parameters (lead) | 323M |
| Variants | 1 |
| Runs with | ggml |
| Released | 2026-07-24 |
| Popularity | 16k downloads / month |
| Likes | 189 |
| Licence | Open weights |
How it works
Variants
Open weights ship in several sizes and precisions. One page, all the variants — pick the one that fits your GPU. VRAM figures are estimates from model size.
| Variant | Params | Precision | VRAM | Fits 16 GB | Weights |
|---|---|---|---|---|---|
| VibeVoice-ASR-BitNet | 323M | BF16 | ~0.7 GB | ✓ | Weights ↗ |
Using it via the API
Once AxForge deploys vibevoice-asr-bitnet for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (vibevoice-asr-bitnet below is illustrative; you get the exact model name on deployment.)
$ curl -sS https://api.axforge.ai/v1/audio/transcriptions \ -H "Authorization: Bearer $AXFORGE_API_KEY" \ -F model="vibevoice-asr-bitnet" -F file=@audio.mp3
Licence
Open weights under mit — commercial use is permitted. Deploy it on AxForge EU hardware on request. Read the licence ↗