I used LLM (gpt o4-mini) for developing a blog template. Overall it did a good job especially with logic loops

Otherwise xml messes up its outputs sometimes

I had to reset my context window frequently

It would do a bad job troubleshooting large sections of code