Artificial Analysis Intelligence Index v5.0

What it measures

Artificial Analysis is an independent evaluator. It runs 9 evaluations itself under one harness and blends them, by four categories, into a single 0–100 index; higher means stronger overall.

The current v5.0 weights each category at 25%: Knowledge Work (AA-Briefcase v1.1, AutomationBench-AA and the long-PDF reasoning test GDP.pdf, which require file deliverables, SaaS workflows driven through APIs or reading long documents); Coding (FrontierCode 2.0, which changes real open-source repositories, and Terminal-Bench 4.0); Science (Terminal-Bench-Science 0.1, research tasks done in a terminal, and the physics-research benchmark CritPt v2 Hard Subset); and General (AA-Omniscience v1.1 for knowledge and hallucination, and HLE-Diamond). It answers “roughly where does this model sit?”

quickly, but it cannot tell you about any single skill. Each reasoning-effort setting (low / high / max) of a model is scored as its own entry.

Reading the scores

a relative 0–100 index, only meaningful among models scored in the same period on the same version. V5.0 is a major revision: it adds FrontierCode 2.0, Terminal-Bench-Science and HLE-Diamond, replaces CritPt with the CritPt v2 Hard Subset, upgrades AA-Omniscience to v1.1, and removes GDPval-AA, SciCode, AA-LCR and Humanity’s Last Exam from the index, so scores should not be compared directly with v4.x. This page includes every entry scoring 45 or above on the default (Current) leaderboard as of November 1, 2026; entries AA shows as -- have no v5.0 index and are left out. Anthropic entries that AA labels “with fallback” include a backup model after refusals, and are named here by reasoning effort.

SourceArtificial Analysis

Read the official Intelligence Index methodology and version history