Last week, NVIDIA quietly unveiled a variation of Meta’s Llama 3.1 model called “Llama-3.1-Nemotron-70B-Instruct.” I didn’t expect the shockwaves it sent through the AI industry, thanks to its superior performance compared to established models like OpenAI’s GPT-4o and Anthropic’s Claude 3.5 Sonnet. The new model’s impressive benchmark scores – 85.0 on Arena Hard, 57.6 on AlpacaEval 2 LC, and 8.98 on GPT-4-Turbo MT-Bench – underscore NVIDIA’s potential to compete in the AI arms race, which is traditionally led by software-centric companies.
NVIDIA’s strategy is clear: by merging hardware prowess with cutting-edge AI software, the company is positioning itself as a full-service AI provider. This new direction not only elevates NVIDIA as a serious AI competitor but also forces industry giants to rethink their approaches. It’s time for the “disrupt the disruptors” phase of our AI journey; of course it’s being brought to us by NVIDIA. You can experiment with the model at build.nvidia.com or on Hugging Face.
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it. This work was created with the assistance of various generative AI models.