Groq has launched its new Llama-3 Tool Use models, specifically the 8B and 70B versions, which have achieved the top position on the Berkeley Function Calling Leaderboard (BFCL). These models are notable for their open-source nature and have been fine-tuned for tool use, demonstrating high performance in function calling tasks. The 70B model achieved an accuracy of 90.76%, while the 8B model reached 89.06%. This release positions Groq's models ahead of established competitors, including OpenAI's GPT-4, Claude, and others. The models are designed to facilitate complex workflows with minimal latency, making them suitable for applications requiring multiple tools and API calls. The launch has been well-received, with contributors from Groq and GlaiveAI recognized for their efforts in developing these state-of-the-art models.