ReDiX/Artemide-3.5

ReDiX/Artemide-3.5 is a 3.82 billion parameter instruction-tuned language model developed by ReDiX, based on Microsoft's Phi-3.5-mini-instruct architecture with a 4096-token context length. It is fine-tuned on a high-quality Italian and English multi-turn conversation dataset, demonstrating strong performance on Italian language benchmarks. This model is optimized for conversational AI tasks requiring proficiency in both Italian and English.