3 ms·
I don't have larger scale experience on this at the moment, but I can tell you what I observed during my trials (also see my related blog entry for some more in
by mpaepper 4y ago
I don't have larger scale experience on this at the moment, but I can tell you what I observed during my trials (also see my related blog entry for some more info: https://www.paepper.com/blog/posts/build-q-and-a-bot-of-your-website-using-langchain/ https://www.paepper.com/blog/posts/build-q-and-a-bot-of-your...)
It cost around 0.05$ to create the embeddings for my ~50 blog entries.
Asking a question in the way I've described it also costs around 0.05$ via the API.
- danielbln 4y ago0.05$ for a question seems expensive, are you using davinci3 or gpt3.5-turbo?