Hacker News Reader

Now we need code quality benchmarks comparing this against GPT-4 and other contenders.

They show the benchmarks in the original post, a few pages down

Thanks, I missed that somehow.