AA-Briefcase Benchmark: Best AI Models Only Fully Solve 3% of Real Knowledge Work
Even the best AI model fails at realistic knowledge work, fully solving just 3 percent of tasks.
Return to original text
A short introduction is retained here; the full text from the source website is not reposted. For specific wording and subsequent updates, please refer to the original article.
Open original at the-decoder.com