I’m sure a lot of people are seeing the debate online about GPT-4o vs Claude 3.5. Many people strongly favour both sides, making it hard to work out which one is best.
What do you guys think?
For my money, I think Claude Sonnet 3.5 is better for domain specific knowledge work. It outperforms GPT-4 on a lot of benchmarks. These benchmarks however are debatably useful, about as good as SATs are at measuring human intelligence-- they miss a lot. Anecdotally, I find it is better at specific subjects like the sciences. Here’s a little reddit thread I was reading about it the other day.
Claude Sonnet 3.5 is much better that for ideating and writing in my opinion.
1 Like