Claude Opus 5.5 tops Artificial Analysis' revised AI index with 58 points
The October 27th update adds coding and scientific-work benchmarks, puts 60% of the score on private or held-out answers, and introduces a separate capability trend index.
By Ryan Merket · Published
Primary source: x.com
Why it matters
Artificial Analysis is making contamination resistance a defining feature of its leaderboard while changing the tests and weights that determine scores. Buyers get a broader comparison that includes task costs, but should treat v5 as a revised measurement system, not a directly comparable continuation of earlier rankings.
Artificial Analysis says 60% of its Intelligence Index v5 now rests on private test sets or held-out answers, as the model evaluator tries to make its leaderboard harder to game. The October 27th release also adds benchmarks for coding and terminal-based scientific work, changes the index's category weights, and introduces a second measure for tracking model capability over time. The company laid out the update in its announcement article and X thread.…
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.