Nous Hermes

Nous Hermes refers to the Hermes series of fine-tuned language models from Nous Research, a lab in the American open source AI movement that works on model architecture, data synthesis, fine-tuning, and reasoning. Hermes models post-train open-weight base models, historically Llama and Mistral among others, on curated instruction data, and the line built its reputation on dependable instruction following, structured outputs, and function calling. The series has continued through Hermes 4, and Nous also publishes agent tooling around the models. Weights are released openly on Hugging Face, so the models run through standard stacks such as Transformers, vLLM, and llama.cpp, and quantized builds circulate for local use. Typical users are developers who want a capable, steerable open assistant model they can self-host, embed in agent pipelines, or fine-tune further without depending on a closed API; licensing follows the respective base model of each release.