import re

sentence= "GPT3, BERT, and T5 are advanced NLP models from OpenAI and Google."
match1 = re.findall(r'\b[A-Z]{2,}[0-9]\b', sentence)
match2 = re.findall(r'[A-Z][a-z]+', sentence)
match3 = re.findall(r'[A-Z].+?[a-z]', sentence)
match4 = re.findall(r'[A-Z][A-Z0-9]+', sentence)
print (match1)
print (match2)
print (match3)
print (match4)