According to BlockBeats, on October 31, OpenAI announced the launch of a new benchmark named SIMPLEQA. This initiative aims to evaluate the factual accuracy of language models. OpenAI has also made this benchmark open-source.