s21mind/HexaMind-Llama-3.1-8B-v25-Generalist

s21mind/HexaMind-Llama-3.1-8B-v25-Generalist is an 8 billion parameter Llama 3.1-based model developed by s21mind, featuring a 32768 token context length. This model excels in reasoning and industrial-grade safety, effectively addressing the "Alignment Tax" by combining strong general intelligence with strict hallucination boundaries. It achieves top-tier performance in math and science reasoning while maintaining high truthfulness, making it suitable for applications requiring both advanced analytical capabilities and robust safety.