VibeVoice is a novel framework designed for generating expressive, long-form, multi-speaker conversational audio, such as podcasts, from text. It addresses significant challenges in traditional Text-to-Speech (TTS) systems, particularly in scalability, speaker consistency, and natural turn-taking. A core innovation of VibeVoice is its use of continuous speech tokenizers (Acoustic and Semantic) operating at an ultra-low frame rate of 7.5 Hz. These tokenizers efficiently preserve audio fidelity while significantly boosting computational efficiency for processing long sequences. VibeVoice employs a next-token diffusion framework, leveraging a Large Language Model (LLM) to understand textual…
VibeVoice is a novel framework designed for generating expressive, long-form, multi-speaker conversational audio, such as podcasts, from text.
Parameters2.7B
Context—
Weights5.4 GB
Licensemit
AccessOpen weights
Monthly Downloads464k
Runs On
What it takes to serve VibeVoice-1.5B (2.7B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 5.4 GB | 6.5 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 2.7 GB | 3.2 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 1.4 GB | 1.6 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.
SAVRN's Notes on VibeVoice-1.5B
Podcast-style audio with several speakers is what Microsoft built VibeVoice for, and its acoustic and semantic tokenizers run at 7.5 Hz. The hardware ask is small: 6.5 GB of memory at 16-bit for 5.4 GB of bfloat16 weights, 3.2 GB at 8-bit, 1.6 GB at 4-bit. The cheapest slot we price is one MI300X with 192 GB at $1.85 an hour on-demand, so the sizing question is how many audio jobs you run in parallel, not whether the model fits.
MIT keeps the paperwork to a copyright and permission notice and permits commercial use, modification and redistribution. The name says 1.5B while the files hold 2.7B parameters, so budget from the 5.4 GB on disk, not the label. No context length is published, so test the script length you plan to synthesize. Nothing prices it per token on the Index, so you run it yourself.
Model Card
By Microsoft, published under mit, revision c00898d257e6.
VibeVoice: A Frontier Open-Source Text-to-Speech Model
VibeVoice is a novel framework designed for generating expressive, long-form, multi-speaker conversational audio, such as podcasts, from text. It addresses significant challenges in traditional Text-to-Speech (TTS) systems, particularly in scalability, speaker consistency, and natural turn-taking.
A core innovation of VibeVoice is its use of continuous speech tokenizers (Acoustic and Semantic) operating at an ultra-low frame rate of 7.5 Hz. These tokenizers efficiently preserve audio fidelity while significantly boosting computational efficiency for processing long sequences. VibeVoice employs a next-token diffusion framework, leveraging a Large Language Model (LLM) to understand textual context and dialogue flow, and a diffusion head to generate high-fidelity acoustic details.
The model can synthesize speech up to 90 minutes long with up to 4 distinct speakers, surpassing the typical 1-2 speaker limits of many prior models.
Technical Report: VibeVoice Technical Report
Project Page: microsoft/VibeVoice
Code: microsoft/VibeVoice-Code
Training Details
Configuration
- Architecture
- VibeVoiceForConditionalGeneration
- Stored precision
- bfloat16
- Model type
- vibevoice
Identity and Version
- Repository
- microsoft/VibeVoice-1.5B
- Publisher
- Microsoft
- Task
- Text to speech
- Modality
- Audio
- Library
- transformers
- Parameters
- 2.7B parameters
- Languages
- en, zh
- Revision
- c00898d257e6b46004e3e2866a47534085fb685a
- First published
- 2025-08-25
- Last updated
- 2026-01-22
Files and Weights
9 files, 5.4 GB in total. The weights are 3 files totalling 5.4 GB in safetensors.
Weights3 files · 5.4 GB
Configuration3 files · 125.7 KB
Documentation1 file · 7.3 KB
Other1 file · 154.0 KB
Repository1 file · 1.6 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00003.safetensors | Weights | 2.0 GB | c5f0a61ddeae |
| model-00002-of-00003.safetensors | Weights | 2.0 GB | 81c3891f7b24 |
| model-00003-of-00003.safetensors | Weights | 1.4 GB | cb6e7e5e86b4 |
| config.json | Configuration | 2.8 KB | — |
| model.safetensors.index.json | Configuration | 122.6 KB | — |
| preprocessor_config.json | Configuration | 351 B | — |
| README.md | Documentation | 7.3 KB | — |
| figures/Fig1.png | Other | 154.0 KB | 64464f28380f |
| .gitattributes | Repository | 1.6 KB | — |
License and Download
- License
- mit
- Access
- Open weights, no gate
- Download size
- 5.4 GB
Download from Microsoft
Released by Microsoft through its official repository on Hugging Face. Read the license.
Built From
- Described by arXiv:2412.08635
- Described by arXiv:2508.19205
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 5.4 GB |
| 16-bit | 5.4 GB |
| 8-bit | 2.7 GB |
| 4-bit | 1.4 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Compare VibeVoice-1.5B
Questions About VibeVoice-1.5B
How much GPU memory does VibeVoice-1.5B need?
About 6.5 GB at 16-bit and 1.6 GB at 4-bit: the weights (2.7B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run VibeVoice-1.5B on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Can I use VibeVoice-1.5B commercially?
Yes. VibeVoice-1.5B is released under MIT License. The MIT License is a short permissive license. It permits commercial use, modification and redistribution, provided the copyright notice and permission notice are included.
Similar Models
https://github.com/vibevoice-community/VibeVoice VibeVoice is a novel framework designed for generating expressive, long-form, multi-speaker conversational audio, such as podcasts, from text. It addresses significant challenges in traditional Text-to-Speech (TTS) systems, particularly in scalability, speaker consistency, and natural turn-taking. A core innovation of VibeVoice is its use of continuous speech tokenizers (Acoustic and Semantic) operating at an ultra-low frame rate of 7.5 Hz. These tokenizers efficiently preserve audio fidelity while significantly boosting computational efficiency for processing long sequences. VibeVoice employs a next-token diffusion framework, leveraging a…
VoxCPM2 is a tokenizer-free, diffusion autoregressive Text-to-Speech model — 2B parameters, 30 languages, 48kHz audio output, trained on over 2 million hours of multilingual speech data. - 30-Language Multilingual — No language tag needed; input text in any supported language directly - Voice Design — Generate a novel voice from a natural-language description alone (gender, age, tone, emotion, pace…); no reference audio required - Controllable Cloning — Clone any voice from a short clip, with optional style guidance to steer emotion, pace, and expression while preserving timbre - Ultimate Cloning — Provide reference audio + its transcript for audio-continuation cloning; every vocal nuance…
MOSS‑TTS Family is an open‑source speech and sound generation model family from MOSI.AI and the OpenMOSS team. It is designed for high‑fidelity, high‑expressiveness, and complex real‑world scenarios, covering stable long‑form speech, multi‑speaker dialogue, voice/character design, environmental sound effects, and real‑time streaming TTS. When a single piece of audio needs to sound like a real person, pronounce every word accurately, switch speaking styles across content, remain stable over tens of minutes, and support dialogue, role‑play, and real‑time interaction, a single TTS model is often not enough. The MOSS‑TTS Family breaks the workflow into five production‑ready models that can be…
svara-TTS is a developer-first multilingual TTS model for 19 languages (18 Indic + Indian English). Built on an Orpheus-style discrete audio token approach, it targets clarity, expressiveness, and low-latency on commodity GPUs/CPUs. It supports light-weight emotion/style control (e.g.,,,, ) and simple speaker identities (Language (Gender)), with zero-shot adaptation paths. Try it live on the Demo Space, or on Colab Deployment scripts and inference repo will be available soon. Watch our Github for updates - Place style/emotion tags at the end of the sentence: आज... सच में अच्छी खबर है — शाम को मिलते हैं! - Use punctuation to hint prosody (ellipses, commas, exclamation). - For technical or…
03/18/2025 – We are releasing our 3B Orpheus TTS model with additional finetunes. Code is available on GitHub: CanopyAI/Orpheus-TTS Orpheus TTS is a state-of-the-art, Llama-based Speech-LLM designed for high-quality, empathetic text-to-speech generation. This model has been finetuned to deliver human-level speech synthesis, achieving exceptional clarity, expressiveness, and real-time streaming performances. Check out our Colab (link to Colab) or GitHub (link to GitHub) on how to run easy inference on our finetuned models. Do not use our models for impersonation without consent, misinformation or deception (including fake news or fraudulent calls), or any illegal or harmful activity. By…