Identity
- Status
- active
- Developer ID
microsoft/VibeVoice-ASR-Streaming-1.5B- Released
- Sep 3, 2026
- Version
- VibeVoice-ASR-Streaming-1.5B
- License
- mit
MicrosoftVibeVoice ASR Streaming 1.5B
VibeVoice-ASR-Streaming-1.5B is an open-weight streaming automatic-speech-recognition model from Microsoft that transcribes speaker-attributed speech, accepts customized hotwords, and supports Chinese, English, French, German, Italian, Japanese, Korean, Portuguese, Russian, and Spanish.
- Token context
- N/A
- Inputs
- Audio
- Outputs
- Text
- Released
- Sep 3, 2026
Providers & Pricing
0 current hosted offerings
No Hosted API Available
This record is downloadable or self-hosted. Its cost depends on the infrastructure and serving configuration, not a universal model price.
At a Glance
Model Facts
Capacity
- Token context
- N/A
- Maximum output
- N/A
- Total parameters
- 2.8B
- Active parameters
- Unknown
- Knowledge cutoff
- Not reported
Interface and Access
- Inputs
- Audio
- Outputs
- Text
- API available
- No
- Open weights
- Yes
- Self-hostable
- Yes
- Reasoning
- No
- Vision
- No
- Tool calling
- No
- Capabilities
- diarization, hotwords, multilingual, streaming, transcription
Technical Specifications
- Supported languages
- Chinese, English, French, German, Italian, Japanese, Korean, Portuguese, Russian, Spanish
- Target sample rate
- 24,000 Hz
- Streaming chunk
- 22 frames
- Streaming lookahead
- 4 frames
This page represents one developer model product. Dated API snapshots, serving endpoints, pricing, regions, quantizations, and service tiers attach as versioned aliases or provider details and never create duplicate public model records.
Version Lineage
PredecessorUnknown
SuccessorsUnknown
Aliasesmicrosoft/VibeVoice-ASR-Streaming-1.5B, vibevoice-asr-streaming-1.5b, VibeVoice ASR Streaming 1.5B
Explicit Unknowns
- Tokenizer
- Knowledge cutoff
Primary Evidence
Sources and Observation Date
microsoft/VibeVoice-ASR-Streaming-1.5B developer model repositoryMicrosoft · observed Sep 2, 2026microsoft/VibeVoice-ASR-Streaming-1.5B Hugging Face inference-provider mappingHugging Face · observed Sep 2, 2026microsoft/VibeVoice-ASR-Streaming-1.5B config.json at 94efa5c0d363Microsoft · observed Sep 2, 2026microsoft/VibeVoice-ASR-Streaming-1.5B README.md at 94efa5c0d363Microsoft · observed Sep 2, 2026microsoft/VibeVoice-ASR-Streaming-1.5B immutable Hub API metadataMicrosoft · observed Sep 2, 2026VibeVoice README at 6df364c86efaMicrosoft · observed Sep 2, 2026microsoft/VibeVoice-ASR-Streaming-1.5B preprocessor_config.json at 94efa5c0d363Microsoft · observed Sep 2, 2026
Comparable Peers
Related Model Comparisons
| A | Pair | B | Context |
|---|---|---|---|
VibeVoice ASR Streaming 1.5BMicrosoft | vs |  Gemini 2.5 FlashGoogle DeepMind | cross-developer peerstext |
VibeVoice ASR Streaming 1.5BMicrosoft | vs | Gemini 2.5 Flash LiteGoogle DeepMind | cross-developer peerstext |
VibeVoice ASR Streaming 1.5BMicrosoft | vs | Gemini 2.5 ProGoogle DeepMind | cross-developer peerstext |
VibeVoice ASR Streaming 1.5BMicrosoft | vs | cross-developer peersaudio, text | |
VibeVoice ASR Streaming 1.5BMicrosoft | vs | cross-developer peersaudio, text | |
VibeVoice ASR Streaming 1.5BMicrosoft | vs |  Muse Spark 1.2Meta | cross-developer peersaudio, text |
VibeVoice ASR Streaming 1.5BMicrosoft | vs | Phi-4 Multimodal InstructMicrosoft | developer peersaudio, text |
VibeVoice ASR Streaming 1.5BMicrosoft | vs | VibeVoice ASR Streaming 7BMicrosoft | family variantsaudio, text |
Questions
VibeVoice ASR Streaming 1.5B FAQs
What is VibeVoice ASR Streaming 1.5B?+
VibeVoice-ASR-Streaming-1.5B is an open-weight streaming automatic-speech-recognition model from Microsoft that transcribes speaker-attributed speech, accepts customized hotwords, and supports Chinese, English, French, German, Italian, Japanese, Korean, Portuguese, Russian, and Spanish. It is developed by Microsoft and its current sourced lifecycle status is active.
What inputs and outputs does VibeVoice ASR Streaming 1.5B support?+
VibeVoice ASR Streaming 1.5B's current record lists Audio as input and Text as output.
What technical specifications are published for VibeVoice ASR Streaming 1.5B?+
VibeVoice ASR Streaming 1.5B's developer-published specifications include supported languages: Chinese, English, French, German, Italian, Japanese, Korean, Portuguese, Russian, and Spanish; target sample rate: 24000 Hz; streaming chunk: 22 frames; streaming lookahead: 4 frames.
Is VibeVoice ASR Streaming 1.5B available through an API?+
VibeVoice ASR Streaming 1.5B is explicitly marked as not API-available in the current record.
How large is VibeVoice ASR Streaming 1.5B?+
The current record lists 2.8B total parameters and an unknown active parameter count. Its architecture is VibeVoiceForASRStreamingTraining.
What capabilities and tasks does VibeVoice ASR Streaming 1.5B support?+
VibeVoice ASR Streaming 1.5B's recorded capabilities are diarization, hotwords, multilingual, streaming, and transcription. Its supported tasks are audio and transcription.
How much does VibeVoice ASR Streaming 1.5B cost through an API?+
VibeVoice ASR Streaming 1.5B has no verified hosted API offering, so no provider token price applies. It is a downloadable or self-hosted artifact whose cost depends on the serving infrastructure and workload.
Can VibeVoice ASR Streaming 1.5B be self-hosted?+
VibeVoice ASR Streaming 1.5B is an open-weight model. Self-hosting is supported, and the recorded license is mit.
Does VibeVoice ASR Streaming 1.5B have related versions or aliases?+
No predecessor is recorded. No successor is recorded. Recorded aliases are microsoft/VibeVoice-ASR-Streaming-1.5B, vibevoice-asr-streaming-1.5b, and VibeVoice ASR Streaming 1.5B.
Which models can VibeVoice ASR Streaming 1.5B be compared with?+
The current compatible peer set includes Gemini 2.5 Flash, Gemini 2.5 Flash Lite, Gemini 2.5 Pro, Granite Speech 5.0 TurboCTC, and Granite Speech 5.0 TurboCTC NC. Compatibility requires shared sourced modalities and tasks; it does not imply that one model is better. Browse model comparisons →