Llama 3 is the latest language model from Meta. It has state of the art performance and a context window of 8000 tokens, double Llama 2's context window.

With Replicate, you can run Llama 3 in the cloud with one line of code.