← Back

Humanity's Last Exam

software 1 mention from 1 sources

A challenging benchmark consisting of 2500 expert-level questions across multiple domains, designed to test advanced AI capabilities.

1

sources

Mentioned by

All mentions

Y Combinator mentioned ✓ High confidence
"So recently you guys just announced some incredible results for humanity's last exam. Can you tell us more about those?"

Attribution: Y Combinator host mentions Humanity's Last Exam as a benchmark Poetic achieved results on