There are no benchmarks that test that. It's my experience and the experience of others who use LLMs for such tasks. Down below there's a comment pushing back

There are no benchmarks that test that. It's my experience and the experience of others who use LLMs for such tasks. Down below there's a comment pushing back on saying Summarization has been solved. Even he/she is saying hallucination is rare.