Instruction Tuned

Alpaca

Stanford's instruction-tuned Llama model — trained on 52k GPT-generated instructions for $600.

Beluga

Stability AI's 2023 Llama fine-tunes — non-commercial, abandoned

Dolly

Databricks' open-source instruction-following model — trained on human-generated data, commercial use permitted.

Nous Hermes

Nous Research's open Hermes models — steerable, minimally filtered

OpenHermes

Teknium's Mistral 7B fine-tune — 2023's go-to local chat model

Zephyr

HuggingFace's instruction-tuned open-source chat model — direct, helpful, and free to use commercially.