Llama 3 70B

The Llama 3 70B model is a state-of-the-art large language model with 70 billion parameters, released by Meta on April 18, 2024. It's based on an auto-regressive transformer architecture and has been optimized for dialogue applications using supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF). The model supports an 8,000-token context length and has been trained on over 15 trillion tokens from public online sources. It excels in tasks such as conversational AI, text generation, and natural language understanding, outperforming many existing open-source chat models on industry benchmarks. The model is designed with a focus on safety and helpfulness, making it suitable for both commercial and research applications, particularly in English. For more details, visit the Hugging Face link .

Llama 3 70B is an open-weight model in the Llama 3 family. The structured metadata tracks a 8k-token context window. This page tracks provider routes through Replicate API, with the cheapest tracked route listed at $0.65 input and $2.75 output per 1M tokens. Headline tracked benchmarks include Google-Proof Q&A 44.1, HellaSwag 92.4, and HumanEval 72.6.