Artificial Analysis launches a 9-test intelligence index that claims to beat every other LLM benchmark
Artificial Analysis Intelligence Index v5.0
Artificial Analysis Intelligence Index v5.0 scores language models across four equally weighted categories—Knowledge Work, Coding, Science, and General—using nine evaluations including AA-Briefcase, Terminal-Bench 4.0, and HLE-Diamond. The suite is text-based and English-only, with multilingual and visual capabilities reported separately. Its makers call it a more useful synthesis than any other metric in existence today, while acknowledging that, like all benchmarks, it has limitations.
We are confident that it is a more useful synthesis comparison between language models than any other metric in existence today.