Skip to main content
ToolPotion
AI Advisor
Submit a Tool (Free)

Apriel Voice 27B Preview

AI ModelsAI Models & LLMsOpen source · Web

Apriel Voice 27B Preview is an open audio-text-to-text model from ServiceNow, built with 27 billion parameters and Transformers. It processes spoken requests, reasons over conversational context, and generates text responses and tool calls. This model enhances interactive voice assistants, customer support, and spoken tool use by adding a speech interface while retaining text and image input capabilities.

View Model
Share
View Model
Share

Apriel Voice 27B Preview is an advanced open model developed by ServiceNow for speech, reasoning, and tool use. This audio-text-to-text model, featuring 27 billion parameters, is built upon the Transformers architecture and leverages a shared Gemma 4 backbone. It is designed to understand spoken requests, engage in reasoning over conversational context, and generate both text responses and structured tool calls.

Read More

The model significantly expands capabilities by integrating a speech interface, while also maintaining its original text and image input functionalities. This makes Apriel Voice 27B Preview highly versatile for developing interactive voice assistants, streamlining customer support workflows, and enabling spoken tool use. When paired with an external speech synthesizer, it can power real-time voice applications.

Key capabilities include comprehensive speech understanding, encompassing transcription, translation, and spoken question answering. It excels in reasoning and tool use, allowing it to follow complex instructions, handle multi-step requests, and generate structured tool calls. The model supports multimodal context, processing audio, text, images, and tool results within a shared conversational framework. For voice applications, it generates text responses that are then paired with external speech synthesis, with application-managed turn-taking.

Technically, Apriel Voice 27B Preview connects a Whisper large-v3-turbo speech encoder to the Gemma 4 26B-A4B backbone via a learned projector. Its training process involves aligning the speech interface and then adapting it with the backbone using speech tasks and text replay. The model supports long audio inputs, automatically splitting 16 kHz mono recordings into 30-second chunks, and has shown good performance on continuous recordings up to 30 minutes, with support for longer durations under specific conditions.

Performance metrics highlight its effectiveness, achieving 51.27% EVA-A, 74.13% EVA-X, and 65.51% τ-Voice task success, with a mean τ-Voice response latency of 1.91 seconds. On LibriSpeech, it demonstrates a word error rate of 2.23% on test-clean and 5.15% on test-other. The model is available for standard and optimized serving via Docker images, providing an OpenAI-compatible API for integration. It is released under an MIT License.

  • CategoryAI Models & LLMs
  • Pricing ModelOpen source
  • DeploymentCloud
  • PlatformsWeb

  • Audio-Text-to-Text Model: Processes spoken requests and generates text responses.

  • 27 Billion Parameters: A large-scale model for complex tasks.

  • Transformers Architecture: Built with a robust and widely used deep learning framework.

  • Speech Understanding: Supports transcription, translation, and spoken question answering.

  • Reasoning and Tool Use: Follows instructions, handles multi-step requests, and generates structured tool calls.

  • Multimodal Context: Integrates audio, text, images, and tool results in conversations.

  • Voice Applications: Generates text responses for external speech synthesis in interactive voice assistants.

  • Gemma 4 Backbone: Utilizes a shared Gemma 4 26B-A4B backbone for core processing.

  • Whisper large-v3-turbo Encoder: Connects speech interface through a learned projector.

  • Long Audio Support: Automatically splits recordings into 30-second chunks, supporting up to 30 minutes continuously.

  • OpenAI-Compatible API: Provides an API for easy integration and inference.

  • Docker Images for Serving: Offers standard and optimized vLLM Docker images for deployment.

  • MIT License: Available for use under an open-source license.

  • High Performance Metrics: Achieves strong results in EVA-A, EVA-X, and τ-Voice task success.

  • Low Word Error Rate: Demonstrates competitive WER on English speech recognition benchmarks.

  1. Access Model: Download or explore the Apriel Voice 27B Preview model files on Hugging Face.

  2. Set Up Environment: Use the provided vLLM Docker image (e.g., `docker.io/amant555/vllm_apriel_voice:plain-26b-v0.32.0`) for serving.

  3. Serve the Model: Run the Docker image on appropriate GPU hardware, exposing an OpenAI-compatible API on port 8000.

  4. Configure Audio Duration: Adjust `VLLM_MAX_AUDIO_DECODE_DURATION_S` and `--max-num-batched-tokens` for desired audio input lengths.

  5. Integrate via API: Send audio or text inputs to the served OpenAI-compatible API using the model name `Apriel-Voice-27B`.

  6. Process Responses: Handle text responses and inspect `response.choices[0].message.tool_calls` for function calling outputs.

  7. Pair with Synthesizer: Integrate an external speech synthesizer for real-time voice interaction in applications.

  8. Optimize Performance: Utilize the optimized Docker image (`docker.io/amant555/vllm_apriel_voice:optimised-26b-v0.32.0`) for higher throughput.

  • Interactive Voice Assistants
  • Customer Support Automation
  • Spoken Tool Use
  • Speech Transcription
  • Speech Translation
  • Spoken Question Answering
  • Multimodal Conversational AI
  • Instruction Following
  • Real-time Voice Interaction

From ServiceNow

Apriel Voice 27B Preview Reviews

Loading...

Popular AI Tools Like Apriel Voice 27B Preview

Gemma 3-27B IT is a state-of-the-art multimodal AI model from Google, capable of handling text and image inputs to generate text outputs. It supports over 140 languages and is…

AI Models & LLMs
View Tool

State-of-the-art multimodal AI model from Google, capable of handling text and image inputs to generate text outputs. It supports over 140 languages and is…Gemma 3-27B IT is a state-of-the-art multimodal AI model from Google, capable of handling text and image inputs to generate text outputs. It supports over 140 languages and is…

AI Models & LLMs
See the Top 5 AI Models

Gemma 3n is a family of lightweight, open-source AI models by Google, designed for efficient execution on low-resource devices. It supports multimodal inputs, including text,…

AI Models & LLMs
View Tool

Family of lightweight, open-source AI models by Google, designed for efficient execution on low-resource devices. It supports multimodal inputs, including text,…Gemma 3n is a family of lightweight, open-source AI models by Google, designed for efficient execution on low-resource devices. It supports multimodal inputs, including text,…

AI Models & LLMs
See the Top 5 AI Models

975B-parameter multimodal AI model designed for developers. It accepts text, image, and audio inputs, generating text outputs for various applications, including…Inkling is a 975B-parameter multimodal AI model designed for developers. It accepts text, image, and audio inputs, generating text outputs for various applications, including…

FeaturedAI Models & LLMs
See the Top 5 AI Models

Gemma 3-4B is a state-of-the-art multimodal AI model from Google, capable of handling text and image inputs to generate text outputs. It supports over 140 languages and is…

AI Models & LLMs
View Tool

State-of-the-art multimodal AI model from Google, capable of handling text and image inputs to generate text outputs. It supports over 140 languages and is…Gemma 3-4B is a state-of-the-art multimodal AI model from Google, capable of handling text and image inputs to generate text outputs. It supports over 140 languages and is…

AI Models & LLMs
See the Top 5 AI Models

Gemma-3n-E4B-it is a lightweight, state-of-the-art AI model from Google, designed for efficient execution on low-resource devices. It supports multimodal input, handling text,…

AI Models & LLMs
View Tool

Lightweight, state-of-the-art AI model from Google, designed for efficient execution on low-resource devices. It supports multimodal input, handling text,…Gemma-3n-E4B-it is a lightweight, state-of-the-art AI model from Google, designed for efficient execution on low-resource devices. It supports multimodal input, handling text,…

AI Models & LLMs
See the Top 5 AI Models

Bark AI Model

AI Models

Bark is a transformer-based text-to-audio model by Suno, capable of generating realistic multilingual speech and other audio forms. It supports research with pretrained model…

AI Models & LLMs
View Tool

AI Models

Transformer-based text-to-audio model by Suno, capable of generating realistic multilingual speech and other audio forms. It supports research with pretrained model…Bark is a transformer-based text-to-audio model by Suno, capable of generating realistic multilingual speech and other audio forms. It supports research with pretrained model…

AI Models & LLMs
See the Top 5 AI Models

Gemma 4 12B Unified Model by Google DeepMind is a multimodal AI model designed for efficient local execution. It supports text, image, and audio inputs, offering advanced…

AI Models & LLMs
View Tool

Gemma 4 12B Unified Model by Google DeepMind is a multimodal AI model designed for efficient local execution. It supports text, image, and audio inputs, offering advanced…

AI Models & LLMs
See the Top 5 AI Models

NVIDIA NemotronLabs VoiceChat is a real-time, full-duplex speech model for conversational AI. It integrates speech understanding and generation in one architecture, reducing…

AI Models & LLMs
View Tool

Real-time, full-duplex speech model for conversational AI. It integrates speech understanding and generation in one architecture, reducing…NVIDIA NemotronLabs VoiceChat is a real-time, full-duplex speech model for conversational AI. It integrates speech understanding and generation in one architecture, reducing…

AI Models & LLMs
See the Top 5 AI Models
1 of 8