// benchmark explainer
Artificial Analysis Intelligence Index, explained
By the TrueIQ editorial team at Ahzy Capital · Last reviewed Oct 15, 2026 · How we work
The Artificial Analysis Intelligence Index is a single number that combines results from a set of evaluations that Artificial Analysis runs itself. News coverage often quotes it when a new model launches.
| Published by | Artificial Analysis (independent) |
|---|---|
| Type | Composite of several benchmarks they run |
| Methodology | artificialanalysis.ai/methodology |
What it measures
A composite averages out the quirks of individual tests, giving a rough overall ranking. Because one organisation runs every model the same way, the numbers are more comparable than scores each lab reports for itself.
Where it falls short
- The mix of benchmarks changes between versions, so an index score from one version isn't comparable with another.
- One number hides strengths and weaknesses. A model can rank lower overall but be best at your task.
- It mostly reflects the kinds of tasks in its component benchmarks.
How to read a score
Use it to shortlist models, then look at the individual benchmarks that match your use case, and at price and speed.
Where to see current scores
We don't republish scores. Leaderboards change weekly, and numbers depend on how the test was run. Check these instead:
- Artificial Analysis: Artificial Analysis Intelligence Index results (independently run)
- Artificial Analysis methodology