Research2026-10-27
Artificial Analysis released Intelligence Index v5. It adds two new benchmarks and upgrades three. For the first time, 60% of the Index uses private datasets. It also launched Progress Index v1 to track capability over time. FrontierCode 2.0 and Terminal-Bench-Science are the new benchmarks. HLE-Diamond, AA-Omniscience v1.1 and CritPt v2 are the upgrades. Claude Opus 5.5 leads the overall Index at 58. The open-weights gap grew from 3 to 12 points. Cost per task varies by more than 13x among the top models. No single model leads all nine evaluations.
What changed
v4 used mostly public benchmarks and a 3-point open-weights gap.
What it unlocks
Comparing current model intelligence on a single composite score.
- 60% of Index now uses private datasets
- Claude Opus 5.5 leads at 58
- Open weights gap widens to 12 points
- Cost per task varies 13x among leaders
Sources