Based on my experience, Chat-GPT4's accuracy in translating English to Chinese is fairly good, at least understandable (though sometimes it looks like gibberish, but not that often).

When translating many text-heavy mods, I have to rely on AI to reduce the workload—first using AI to translate, then checking for awkward parts and correcting them with reference to the original text, and finally reading and revising the whole thing again.

Before using it, I use many sentences to make GPT's answers as close to my requirements as possible. The structured sentences I use are:
Then I send what I want to translate.
Example 1
I sent him this content:
Then GPT returns this content to me:
In this case, the translation is very good. I didn't make any modifications and used this part directly.

Example 2:
At this point, GPT returned:
This result didn't satisfy me very much, but it's still understandable and there's no major problem. Overall, GPT-4's translation is mostly understandable, occasionally unsatisfactory, and in rare cases perfect on the first try (Example 1).