Nvidia recently launched an open-source AI model called Llama-3.1-Nemotron-70B-Instruct, which reportedly outperforms leading AI systems like GPT-4o and Claude-3. The model is a modified version of Meta's Llama-3.1-70B-Instruct, enhanced by Nvidia using curated datasets and advanced fine-tuning techniques to improve its helpfulness. The developers claimed that in comparative testing on the Chatbot Arena, Nemotron achieved an impressive score of 85, positioning it as a top performer, despite being built on a mid-tier model. In contrast, GPT-4o is believed to have over one trillion parameters, further highlighting the significance of Nemotron’s performance based on Meta's smaller model. The ongoing debate on AI performance evaluation underscores the subjectivity in determining the 'best' model, relying on comparative testing to assess their efficacy.

Source 🔗