Nvidia's Llama-3.1-Nemotron-70B-Instruct model has emerged as a leading open-source large language model (LLM), excelling in multiple benchmarks. It scored 85 on the Arena Hard benchmark and 55% on aider's leaderboard, just behind the plain Llama-3.1-70B-Instruct, which scored 59%. The model is noted for its predictive ability in Chatbot Arena Elo scores. It has been praised for its performance, surpassing other models like Sonnet 3.5 and GPT-4o. The model is available on various platforms, including Baseten's model library and Nvidia's API catalog. The Llama-3.1-Nemotron-70B-Instruct model also includes the 3 Rs in Strawberry feature.