Model reference · open weights
VibeVoice-ASR-BitNet is an open-weight audio or speech model from microsoft. VibeVoice-ASR-BitNet (FP32) weighs 993 MB; the smallest configuration that runs it is RTX 3060 12 GB.
Summary of the microsoft/VibeVoice-ASR-BitNet model card, 2026-10-01
What it is
| Released by | microsoft |
|---|---|
| Released | 2026-07-24 |
| Parameters | 323M |
| VRAM | 993 MB for the weights |
What it runs on
| Card | One stream | Memory |
|---|---|---|
| RTX 3060 12 GB | fits | 11.6 GB |
| RTX 4060 Ti 16 GB | fits | 15.4 GB |
| RTX 3090 24 GB | fits | 23.4 GB |
| RTX 4090 24 GB | fits | 23.4 GB |
| RTX 5090 32 GB | fits | 31.0 GB |
| L40S 48 GB | fits | 44.0 GB |
| A100 80 GB | fits | 78.2 GB |
| H100 80 GB | fits | 78.1 GB |
| RTX PRO 6000 Blackwell 96 GB | fits | 93.8 GB |
| DGX Spark (GB10) 128 GB unified | fits | 107 GB |
| H200 141 GB | fits | 138 GB |
| B200 180 GB | fits | 176 GB |
From the model card
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 ➡️ Report: VibeVoice-ASR-BitNet Technical Report ➡️ Base Model: microsoft/VibeVoice-ASR
| Component | FP16 | Quantized | Method | Compression |
|---|---|---|---|---|
| VAE Tokenizer | 1.31 GB | 0.65 GB | I8_S | 2.0× |
| LM Decoder | 3.32 GB | 0.92 GB | I2_S + Q6_K | 3.6× |
| Total | 4.62 GB | 1.58 GB | — | 2.9× |
| Threads | 1 | 2 | 3 | 4 | 6 | 8 |
|---|---|---|---|---|---|---|
| RTF | 1.98 | 1.08 | 0.77 | 0.63 | 0.49 | 0.42 |
| vs. Whisper.cpp | 2.28× | 2.12× | 1.86× | 1.86× | 1.71× | 1.55× |
Benchmarked on AMD EPYC 7V13 (AVX2+FMA) with 20s audio. Bold = RTF < 1 (real-time).
| Benchmark | VibeVoice-ASR-7B | VibeVoice-ASR-BitNet | Parakeet | Whisper | SenseVoice | FunASR |
|---|---|---|---|---|---|---|
| MLC-EN | 7.82 | 8.25 | 8.40 | 13.57 | 12.39 | 11.36 |
| MLC-FR | 16.03 | 17.41 | — | — | — | — |
| MLC-IT | 15.67 | 17.23 | — | — | — | — |
| MLC-KO | 9.83 | 11.15 | — | — | — | — |
| MLC-PT | 22.41 | 24.87 | — | — | — | — |
| MLC-VI | 20.15 | 22.38 | — | — | — | — |
| AISHELL4 | 19.83 | 27.45 | — | — | 22.52 | 20.41 |
| AMI-ihm | 17.42 | 21.36 | 21.92 | 27.07 | 30.81 | 32.07 |
| AMI-sdm | 24.18 | 25.87 | 26.33 | 36.92 | 48.11 | 40.17 |
| AliMeeting | 36.21 | 40.58 | — | — | 38.75 | 39.27 |
| Fleurs-en | 4.73 | 5.21 | 4.09 | 3.99 | 6.84 | 4.93 |
| Fleurs-zh | 7.92 | 8.35 | — | — | 5.56 | 7.00 |
| Libri-clean | 2.17 | 2.41 | 1.49 | 1.98 | 2.78 | 1.58 |
| Libri-other | 5.84 | 6.27 | 3.13 | 3.60 | 6.81 | 4.01 |
| VoxPopuli | 4.92 | 5.18 | 5.26 | 7.19 | 8.63 | 6.46 |
| File | Size | Description |
|---|---|---|
vibeasr-vae-encoder-i8_s.gguf | 0.65 GB | VAE tokenizer, I8_S quantized (ready to use) |
vibeasr-lm-i2_s-embed-q6_k.gguf | 0.92 GB | LM decoder, I2_S quantized (ready to use) |
model-*.safetensors | 10.7 GB | Original SafeTensors (for conversion) |
This project is licensed under the MIT License.
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.
Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.