4 ms·

Given the high price due to high token count, I wonder how different the results would be running the same analysis with a local model.