LFM3: six small open-weight models from Liquid AI
Models in the 2B-parameter class for text, vision, audio, omni and decisions. The cards list Greek only among speech languages.
Source of truthhuggingface.co/blog/LiquidAI/lfm3On October 27 Liquid AI released LFM3: six models in the 2B-parameter class on a shared backbone, for text, images and video, audio, an omni model that combines them, and two decision models, d1-2B and d1-omni-2B, for calibrated answers. LFM3-1.7B (text) has a 256K-token context and, according to the company, needs about 1 GB of memory at 4-bit. The model cards list Greek only among the speech languages of Audio-2B and Omni-2B; the text-language list of LFM3-1.7B does not include it. The cards do not say whether Greek text is supported and give no Greek benchmarks. The LFM Open License v1.0 of LFM3-1.7B does not cover commercial use by companies with annual revenue of $10 million or more. The LFM3-1.7B card does not recommend it for agentic coding or complex multi-step reasoning.
Other sources: huggingface.co/LiquidAI/LFM3-1.7Bhuggingface.co/LiquidAI/LFM3-VL-2Bhuggingface.co/LiquidAI/LFM3-Audio-2Bhuggingface.co/LiquidAI/LFM3-Omni-2Bhuggingface.co/…/LICENSE
Related stories
Mistral AI releases the open weights of Mistral Large 4
The weights of the model, with about 1 trillion parameters, can be downloaded from Hugging Face under a modified MIT license. On the API it costs $0.70/$2.10 per million tokens.
Why it matters → For a Greek software company or an accounting firm building its own tools, one option is the API: $0.70 per million input tokens and $2.10 for output. If you want to sell it as a service, you need a commercial license. Mistral AI says the model is fluent in every official EU language, so Greek too, but gives no figures on its quality in Greek.
Replica: GitHub's open-source AI prototyping tool, under MIT
According to GitHub, Replica builds interactive prototypes from your team's design system, through the GitHub Copilot app.