Research scope: This study is a snapshot: the 21 evaluations listed on the Artificial Analysis Intelligence Benchmarking methodology page on November 3, 2026, coded by Tokely to the 41 Generalized Work Activities in O*NET 31.0. It measures which work activities the evaluations test, not how well any model performs them, and occupations are not weighted by employment.
Short answer: The Artificial Analysis Intelligence Index tests 8 of 41 O*NET work activities and puts 73% of its weight on three of them. Across all 21 evaluations listed on its methodology page, coverage widens to 13 of 41 activities. This leaves what we call the colleague gap, where communicating with supervisors, peers or subordinates is rated very important (4 or higher on O*NET's 1 to 5 importance scale) in 541 of 911 occupations, yet 0 of 21 evaluations test it.
What did we measure?
The US Department of Labor catalogs occupational requirements across the economy using the O*NET 31.0 database.…