Has anyone tried with open source LLM or gemini free usage?

[Isaadahmed2]

Is there is anyway we can use hugging face model or open source LLM or quantized Open sources LLM for the purpose or we can use gemini free usage quota?

[MarlNox]

I've tried it many times, it seems it only works properly with GPT4o for the moment. I assume a finetune of the local LLMs, seems to be the way to go from here.

[LMJOK]

I'm trying, but my localization of the large model has been calling unsuccessfully

[xprabhudayal]

try llama3.1 api from groq its free, Added support for Groq

[Masao-Taketani]

@Isaadahmed2 @LMJOK Big thanks to the authors of this repository, I just have created a repository that accommodates OSS local models to back up the AI Scientist! Please check it out! URL: https://github.com/Masao-Taketani/AI-Scientist-with-Local-LLMs

It currently supports Hugging Face or Ollama platforms including quantized models, and it even enables you to use the recently-announced DeepSeek R1 models!

May the AI Scientist Locally be with You!