From my tests, GLM tends to be best for server code or frontend logic. It's not very good at design tasks. It did make a killer chess app with UX, but I think i

From my tests, GLM tends to be best for server code or frontend logic. It's not very good at design tasks. It did make a killer chess app with UX, but I think it was just trained heavily for it.