Reported in recent model launches
These model reports include this benchmark. Their harnesses, task subsets, scoring policies, and resource budgets can differ; follow each report’s evaluation notes before joining the numbers.
What changed in v5.0
The current suite adds AutomationBench-AA, FrontierCode 2.0, HLE-Diamond, Terminal-Bench-Science and CritPt v2’s hard subset. GDPval-AA, SciCode, AA-LCR and original HLE and CritPt results remain useful standalone evidence.
Reading an index score
Use the component leaderboards to find where models differ. A composite depends on its task mix and weights; the score alone does not identify a suitable model for a particular workflow.
Historical comparisons
Select an index version before comparing scores. Older reports are retained with their original dates and declared versions. An unspecified report version is not treated as v5.0.
Where this fits in an evaluation suite
Ask what the task actually requires: recall, deduction, search, or a mixture. Separate closed-book reasoning from retrieval and tool-assisted problem solving.
Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.
A reproducible comparison
Published v5.0 scores; four categories weighted 25% each. Retain effort settings and keep every earlier index revision separate.
Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.
PRIMARY EVIDENCE & FURTHER READING
- Artificial Analysis Intelligence Index v5.0 — Artificial Analysis ↗
- Artificial Analysis Intelligence Benchmarking Methodology — v5.0 ↗
Reviewed 2026-10-28. Explanation and interpretation by RLTasks. Benchmark ownership remains with its original authors.