LFM3-Omni-2B is Liquid AI's first omni-modal model in the LFM3 family: a single end-to-end model that understands text, images, video and audio, and answers in text and speech. It is built for on-device use: one model replaces a separate ASR model, vision-language model, LLM and TTS model.
Alternatives to LFM3-Omni-2B
114 open-weight models that do the same job as LFM3-Omni-2B (2.2B omni), ranked by how close they are in task and size, then by popularity and how recent they are. Licence and VRAM are shown for each.
Ming-UniAudio is a novel framework that unifies speech understanding, generation, and editing. Its core is a unified continuous speech tokenizer that effectively unifies semantic and acoustic features within an end-to-end model.
18.3BA3B32K ctx Apache 2.0171 80Sept 2025- Same category
- Larger (18.3B vs 2.2B)
- Commercial use OK
If you find TeleStyle useful for your research and applications, please cite using this BibTeX:
?BApache 2.00 64Jan 2026- Same category
- Commercial use OK
The public training-content summary identifying Ming-UniVision-16B-A3B is available below. Please refer to the document for its covered model versions, training-content scope, summary version, and update date.
18.8BA3B32K ctx Apache 2.0187 65Sept 2025- Same category
- Larger (18.8B vs 2.2B)
- Commercial use OK
For more features and to experience the full capabilities of our product, please visit https://vivago.ai/.
17.1BMIT218 217Apr 2025- Same category
- Larger (17.1B vs 2.2B)
- Commercial use OK
🔔 Latest: Emu3.5 Web & Mobile Apps and vLLM offline inference are live — see 🔥 News for details.
34.1B32K ctx Apache 2.0399 176Oct 2025- Same category
- Larger (34.1B vs 2.2B)
- Commercial use OK
For more features and to experience the full capabilities of our product, please visit https://vivago.ai/.
17.1BMIT119 217Jul 2025- Same category
- Larger (17.1B vs 2.2B)
- Commercial use OK
RealVideo is a WebSocket-based video calling system that supports text input. It leverages GLM-4.5-AirX and GLM-TTS models to generate audio responses and utilizes autoregressive diffusion to generate corresponding video frames.
?BMIT0 107Dec 2025- Same category
- Commercial use OK
We introduce LongCat-Flash-Omni, a state-of-the-art open-source omni-modal model with 560 billion parameters (with 27B activated), excelling at real-time audio-visual interaction, which is attained by leveraging LongCat-Flash's high-performance Shortcut-connected Mixture-of-Experts (MoE)…
561B128K ctx MIT322 115Oct 2025- Same category
- Larger (561B vs 2.2B)
- Commercial use OK
OmniLMM-12B is the most capable version of OmniLMM currently. The model is built based on EVA02-5B and Zephyr-7B-β, connected with a perceiver resampler layer, and trained on multimodal data in a curriculum fashion. The model has three notable features:
12B32K ctx Unknown licence178 74Jan 2024- Same category
- Larger (12B vs 2.2B)
Step-Audio-R1 is the first audio language model to successfully unlock Chain-of-Thought (CoT) reasoning. It decisively solves the "inverted scaling" problem that plagues existing models, where performance degrades with longer reasoning.
33.5BApache 2.0119 147Nov 2025- Same category
- Larger (33.5B vs 2.2B)
- Commercial use OK
A Multi-Codebook Diffusion Large Language Model for Omni-Visual Understanding, Image Generation and Editing.
16BApache 2.00 23Sept 2026- Same category
- Larger (16B vs 2.2B)
- Commercial use OK
This repository contains the Multimodal Large Language Model (LLM) component of Step-Audio. It is a 130 billion parameter multimodal LLM that is responsible for understanding and generating human speech.
132B32K ctx Apache 2.0267 461Feb 2025- Same category
- Larger (132B vs 2.2B)
- Commercial use OK
UniPic2-SD3.5M-Kontext-2B is a post-trained T2I model built on the SD3.5-Medium. It focuses on text-to-image generation and image editing, delivering strong quality with a fast generation speed. It runs smoothly on a single 16 GB consumer GPU.
5MMIT69 25Aug 2025- Same category
- Smaller (5M vs 2.2B)
- Commercial use OK
MOSS-Transcribe-Diarize 0.9B is an end-to-end audio understanding model for long-form multi-speaker transcription, diarization, timestamps, and acoustic event awareness.
909MApache 2.0148.9K 528May 2026- Related category
- Smaller (909M vs 2.2B)
- Commercial use OK
VibeVoice-ASR is a unified speech-to-text model designed to handle 60-minute long-form audio in a single pass, generating structured transcriptions containing Who (Speaker), When (Timestamps), and What (Content), with support for Customized Hotwords and over 50 languages.
8.3BMIT16.4K 190Mar 2026- Related category
- Larger (8.3B vs 2.2B)
- Commercial use OK
ACE-Step Transcriber is the annotation model used by ACE-Step v1.5 for training data labeling. It is a powerful multilingual audio transcription model capable of transcribing both speech and singing voice with high accuracy.
10.7BMIT1.8K 66Jan 2026- Related category
- Larger (10.7B vs 2.2B)
- Commercial use OK
OLMoASR is a series of English automatic speech recognition (ASR) models proposed in the OLMoASR: Open Models and Data for Training Robust Speech Recognition Models paper by Huong Ngo et al. from Ai2.
?BApache 2.00 78Jul 2025- Related category
- Commercial use OK
Ichigo Whisper is a compact (22M parameters), open-source speech tokenizer for the Whisper-medium model, designed to enhance performance on multilingual with minimal impact on its original English capabilities.
22MCC BY-NC-SA 4.00 28Dec 2024- Related category
- Smaller (22M vs 2.2B)