No, not fully. ChatGPT and Google Translate/DeepL are about equal. This is for a couple reasons: machine translations are very good these days, and ChatGPT struggles with domain specific translations (whereas these commercial products have incentive to improve within).
It's marginally better at sounding natural with translations though. "In fact, the frequent use of soap is bad for the skin." (chatgpt) vs "In fact, using soap too often is bad for your skin." (google)[1]
I know what the paper says but actually using both, GPT-4 is far ahead of google and Deepl. I think the isolated one sentence datasets used for evaluations are no longer up to snuff.
Trying something longer and more comprehensive and the difference is very clear.
There is no comparison in my experience, especially when going between wildly unrelated languages instead of Latin-derived European languages.
I tested some complex sentences translated from English to Hungarian to Chinese and then back to English and it preserved 90-95% of the meaning. Google Translate will shred anything that goes through those kinds of steps.
It can also translate to Latin, which shouldn't be surprising but was still a bit of a shock when I tried it. It's Elvish and Klingon is not so good however, the round-trip only vaguely preserves the meaning.
Maybe not quite ChatGPT 2023 but it looks like from here it's just a matter of refinement. Someone may always create a better toothpaste, but I don't expect to ever really be surprised and impressed by a new toothpaste. ChatGPT is in the same place as a translator.