unsloth/Llama-3.1-8B-Instruct

The unsloth/Llama-3.1-8B-Instruct is an 8 billion parameter instruction-tuned generative language model developed by Meta, part of the Llama 3.1 collection. Optimized for multilingual dialogue, it utilizes an optimized transformer architecture with Grouped-Query Attention and a 128k context length. This model excels in assistant-like chat applications and supports tool use, demonstrating strong performance across general, reasoning, code, and math benchmarks.