Has anyone tried both Ollama and LMStudio?

[JerryZLiu]

Has anyone tried both LMStudio and Ollama. For some reason I'm seeing some drastically differences. Ollama takes about 5 minutes to process a batch, whereas LMStudio only takes about 40 seconds. I don't know enough about how each works under the hood, but this seems pretty drastic.

Has anyone observed anything similar?

[brooksc]

LM Studio supports MLX which means LLMs require less resources to run. Ollama doesn't. I stopped using Ollama when I learned about that. Is there a reason you want to use ollama?

[JerryZLiu]

Makes sense, only reason why I supported Ollama was that I thought it was the most popular way to run llms locally, at least according to github stars.

[brooksc]

Sorry Jerry I missed that you were asking as a product author. Yes supporting ollama makes sense -- I'd just encourage anyone on a Mac to use LM Studio for better performance.