VibeVoice-ASR Streaming 1.5B vs VibeVoice-ASR-Streaming-7B
Two sizes of Microsoft's native streaming who-said-what model. VibeVoice-ASR Streaming 1.5B ranks higher in all seven of the matcher's common setups, but the right choice still depends on the differences below.
Key differences
- VibeVoice-ASR Streaming 1.5B is the lighter option (memory rating 2/5 versus 1/5).
- VibeVoice-ASR Streaming 1.5B carries the higher relative speed rating (4/5 versus 3/5).
- VibeVoice-ASR Streaming 1.5B carries the higher relative accuracy rating (4/5 versus 3/5); confirm it on your own audio.
- VibeVoice-ASR Streaming 1.5B is rated better for Apple Silicon (2/5 versus 1/5).
- The main documented runtime is microsoft/VibeVoice for VibeVoice-ASR Streaming 1.5B and Hugging Face Transformers for VibeVoice-ASR-Streaming-7B.
Side by side
| Fact | VibeVoice-ASR Streaming 1.5B | VibeVoice-ASR-Streaming-7B |
|---|---|---|
| Provider · size | Microsoft · 3B total | Microsoft · 9B total |
| License | MIT Permissive | MIT Permissive |
| Languages | 10 listed languages | 10 listed languages |
| Streaming | Native | Native |
| Timestamps | Limited | Limited |
| Speaker diarization | Native | Native |
| CPU · Apple · NVIDIA fit | 1/5 · 2/5 · 4/5 | 1/5 · 1/5 · 4/5 |
| Live · long recordings | 5/5 · 4/5 | 4/5 · 4/5 |
| Main runtime | microsoft/VibeVoice | Hugging Face Transformers |
Which ranks higher in common setups
Scores from the same deterministic matcher as the homepage, applied to all 26 reviewed models (rank out of 26). They are decision guidance, not accuracy claims.
| Setup | VibeVoice-ASR Streaming 1.5B | VibeVoice-ASR-Streaming-7B | Higher |
|---|---|---|---|
| Long recordings on an 8–12 GB NVIDIA GPU, balanced | 68/100 · #15 | 65/100 · #24 | VibeVoice-ASR Streaming 1.5B |
| Live captions or voice input (native streaming + timestamps, speed first) | 83/100 · #5 | 77/100 · #14 | VibeVoice-ASR Streaming 1.5B |
| Subtitles with timing, accuracy first | 67/100 · #18 | 61/100 · #23 | VibeVoice-ASR Streaming 1.5B |
| Meetings with speaker labels (diarization + timestamps) | 77/100 · #9 | 74/100 · #13 | VibeVoice-ASR Streaming 1.5B |
| CPU-only machine, low memory use | 50/100 · #23 | 48/100 · #24 | VibeVoice-ASR Streaming 1.5B |
| Apple Silicon with 8–16 GB unified memory, long recordings | 58/100 · #23 | 50/100 · #26 | VibeVoice-ASR Streaming 1.5B |
| Japanese live transcription (native streaming, speed first) | 83/100 · #3 | 77/100 · #5 | VibeVoice-ASR Streaming 1.5B |
Language coverage
Both list 10 languages: English, French, German, Italian, Japanese, Korean, Mandarin Chinese, Portuguese, Russian and Spanish. VibeVoice-ASR Streaming 1.5B lists no language that VibeVoice-ASR-Streaming-7B lacks. VibeVoice-ASR-Streaming-7B lists no language that VibeVoice-ASR Streaming 1.5B lacks.
How to run each
- VibeVoice-ASR Streaming 1.5B
- microsoft/VibeVoice, vLLM streaming server.
git clone https://github.com/microsoft/VibeVoice.git cd VibeVoice pip install -e .Microsoft's documentation uses an NVIDIA GPU container.
- VibeVoice-ASR-Streaming-7B
- Hugging Face Transformers, microsoft/VibeVoice.
pip install -U transformers accelerateNo memory figure is published; the card lists 9B parameters.
Runtimes and figures from the official sources checked on November 2, 2026; details on each VibeVoice-ASR Streaming 1.5B and VibeVoice-ASR-Streaming-7B profile.
Common questions
Is VibeVoice-ASR Streaming 1.5B better than VibeVoice-ASR-Streaming-7B?
It depends on the job. VibeVoice-ASR Streaming 1.5B ranks higher in all seven of the matcher's common setups, but the right choice still depends on the differences below. Benchmark both on your own audio before deciding.
Which is lighter to run, VibeVoice-ASR Streaming 1.5B or VibeVoice-ASR-Streaming-7B?
VibeVoice-ASR Streaming 1.5B has the lighter memory rating (2/5 versus 1/5). VibeVoice-ASR Streaming 1.5B: Microsoft's documentation uses an NVIDIA GPU container. VibeVoice-ASR-Streaming-7B: No memory figure is published; the card lists 9B parameters.
Which supports more languages?
Both list 10 languages in our record.
Go further
- Add a third model in the interactive comparison
- Benchmark both on your own audio
- All model comparisons
Generated from the reviewed model database (latest review November 3, 2026). Differences are taken from primary sources; matcher scores are Speech Model Match's editorial decision scores.