The benchmark is called "random circuit sampling". https://blog.google/technology/research/google-willow-quantu... I call it a benchmark because a random circuit is probably not useful, but quantum computers are useful for simulating quantum systems including materials science, particle physics etc. This result is separate from the error-corrected qubit result and did not use error correction. There is a third result in the paper, which is a repetition-coded qubit that was stable for almost an hour. https://www.nature.com/articles/s41586-024-08449-y