The arrival of the 86B model signals a crucial shift in the field of large language models. Its impressively smaller scale compared to behemoths like GPT-4 shouldn't necessarily indicate weakness, but 86blegal.com