Independent analysis of AI models
QBench compares models on public benchmark results. The QBench Composite Index 1.0 keeps the four-category structure and evaluation criteria of the Artificial Analysis Intelligence Index v5 as a reference. No Artificial Analysis values are used.
QBench Composite Index 1.0 leaderboard
QBench Index vs. API price
Equal-token list-price proxy (USD per 1M tokens, log scale). Not a measured cost.
Category breakdown, top ranked models
Each category contributes up to 25 points of the index.
Models
Every model with at least two imported results. Select up to five to compare.
Comparison
Value (percentile within benchmark). Missing results stay blank.
Benchmarks
Select a benchmark to see its model scores. The status shows whether it enters the index.