I'm not sure. They even released tool use benchmarks for Llama 3.1

→ View original post on X — @maximelabonne