Mistral-7B OpenOrca is a good model to start with - https://huggingface.co/Open-Orca/Mistral-7B-OpenOrca
See this example on how to run this model: https://neuml.hashnode.dev/build-rag-pipelines-with-txtai. Plenty of other models (https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderb...) that may work better depending on the use case.