Pete uses a large language model to work. There are many to choose from and initially I had set up Pete to use the Perplexity model which is US based. This is obviously counter to what the site is all about so I have been investigating other models. The site now uses Alibaba’s Qwen model (Chinese).
Now there are obvious issues with using a Chinese model instead but the French Mistral models didn’t seem to work so well in testing. I’ll continue to look but aside from the US and China there aren’t any widely available models.
The second reason to go for Qwen for the time being is cost. I’m funding this at present and using an AI to answer these queries costs money. It costs significantly less using Qwen. If you would like to make a donation please go here.

Ive done some testing for this change but if you would like to give any feedback id love to hear it here.