Llama-3.1-405B

Llama 3.1 405B is Meta's flagship open-weight foundation model, released in July 2024 as part of the Llama 3.1 family alongside 8B and 70B variants. With 405 billion parameters, it was the largest openly available LLM at launch and is designed to compete with leading closed models such as GPT-4 and Claude 3.5 Sonnet across reasoning, coding, mathematics, and multilingual tasks.

Built on the Llama 3 decoder-only Transformer architecture with Grouped-Query Attention (GQA), the model supports a 128K token context window and eight languages including English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai. It was pretrained on roughly 15 trillion tokens of publicly available data and demonstrates strong performance on MMLU (85.2 base / 88.6 instruct CoT), GSM8K (96.8), HumanEval (89.0), and MATH (73.8–73.9).

Llama 3.1 405B is released under the Llama 3.1 Community License for research and commercial use. Due to its scale, full-precision inference requires approximately 810 GB of VRAM, making multi-GPU or multi-node deployments standard. Quantized variants (FP8, INT4) and companion instruct-tuned, FP8, and tool-use models extend its practical deployment options.