BENCHMARK GUIDE
What the Artificial Analysis Intelligence Index measures
It combines multiple evaluations into an overall index. The benchmark mix, weights and grading methods can change between versions.
When it is useful
Use it to survey performance across several kinds of task. Record the index version and the exact model configuration alongside the value.
What it cannot tell you
Do not join an old index score and a new version’s score into one growth curve without a compatible recalculation. The methodology’s version history documents substantial changes.
If one score uses a different exam mix, the difference can reflect both a stronger model and a changed test. Keep the two explanations separate.
The chart shows selected published results with a source, snapshot date and tested configuration. Coverage differs by benchmark. Missing scores stay blank; demo values are no longer shown.
Sources
Developer reports describe their own results. This site has not independently reproduced them.
Artificial Analysis methodology and version history