This is an MLX export of the LFM2.5-1.2B-Instruct model optimized for Apple Silicon inference. The model is a 1.2 billion parameter instruction-tuned language model quantized to 4-bit precision, resulting in a 628 MB footprint while maintaining a 128K context length. It supports multiple languages including English, Japanese, Korean, French, Spanish, German, Italian, Portuguese, Arabic, and Chinese, making it suitable for multilingual text generation tasks on edge devices.