LFM3-Audio-2B
Liquid AI’s official 2B speech model from the LFM3 family: speech recognition, text-to-speech and spoken chat in one small model that runs on a laptop.
What do you want to do?
Choose a task. Get the software, files, settings, and steps to run it.
Open a guide for its software, hardware requirements, and setup steps.
At a glance
- Parameters
- 2B
- License
- LFM Open License v1.0
- Software
- llama.cpp, Ollama, MLX, Transformers, vLLM
Will it run on your machine?
Save your machine to see a personalized rating and its reasoning.
Add my hardwareBest for
- Transcribe recordings privately.
- Read text aloud into an audio file.
Tradeoffs
- Two built-in voices (male and female).
- Realtime voice chat in llama.cpp is experimental.
- This is the official audio model; the site’s LFM3-1.7B Uncensored page is a separate community text model.
Ways to run it
llama.cpp v0.8.0+ · Windows, macOS, Linux
Publisher GGUF and commands.
llama-cli -hf LiquidAI/LFM3-Audio-2B-GGUF:Q4_K_M --audio speech.wav -sys "Perform ASR in English." -p "<__media__>" --temp 0What to expect
Use the Q4_K_M GGUF with llama.cpp v0.8.0 or newer for private transcription and speech on any desktop.
Explore its uses
Sources and files
Primary sources and files behind this model page.
Keep exploring
Voxtral Mini · 4B Realtime
Voxtral Mini 4B Realtime 2602 is a multilingual, realtime automatic speech recognition model from Mistral AI, released under the Apache-2.0 license. It features a natively streaming architecture with a causal audio encoder and sliding window attention, enabling configurable transcription delays from 240ms to 2.4s (and reportedly 80ms to 2.4s). The model has approximately 4B parameters (≈3.4B language model and ≈970M audio encoder) and is described as optimized for on-device deployment. It supports 13 languages and is presented as achieving accuracy comparable to offline systems with less than 500ms delay, specifically matching leading offline open-source models and realtime APIs at a 480ms delay. The model card recommends a temperature of 0.0, using websockets for audio streaming, and a delay of 480ms as a balance between performance and latency.
YuE2 · 3B
Create a full song from style and lyrics with a publisher Python package on a Linux NVIDIA GPU; outputs include a playable FLAC.
granite speech 5.0 470m turboctc
Granite Speech 5.0 470M TurboCTC is a compact 470 million parameter English automatic speech recognition model released by IBM under the Apache 2.0 license. It uses a conformer acoustic encoder with block self-attention, self-conditioning, and temporal downsampling, with an output layer of 16,384 BPE units. The model was trained on approximately 60,000 hours of English audio from public corpora using Connectionist Temporal Classification (CTC), and inference is non-autoregressive with greedy decoding. It is designed for low-latency and high-throughput English speech-to-text transcription, and is described as suitable for deployment on laptops, smartphones, and other edge devices. Native support is available in transformers>=5.16.0, with additional usage options for Apple Silicon via mlx-audio and local GGUF inference via transcribe.cpp.