It seems on certain result, some text data when relating to maths does not seem to be outputting correctly.
This was on a "Customised" asisstant based on GPT 4 Turbo.
Example conversation: https://kagi.com/assistant?mode=6&sub_mode=NaN&thread=y6erghD0BzhJBl5uZCieDl6isaLavRLR
Screenshot from the relevant section:
Text formatting should take effect rather than the raw characters being printed.
Vlad I've tried this now, it looks like it formats it correctly in what looks to be a different text font and central to the prompt output.
I assume this is the desired effect, as no formatting code is printed raw in the response anymore.
It is interesting also as it seems the Claude 3 model does not seem to prefer using formatting in this way, but GPT-4-Turbo does. 
I seem to still come across similar rendering problem occasionally at least with gpt 4 turbo. Example conversation: https://kagi.com/assistant?mode=4&sub_mode=7&thread=Qv2HLl2PFhdskg0E86qmGytUN8F88h8u
Screenshot:
