Artificial Analysis's composite score combining performance across multiple benchmarks into a single intelligence metric.
Provides a single comparison point by aggregating multiple benchmarks into one ranking signal; reduces volatility versus relying on any single benchmark snapshot.
Weighting is set by Artificial Analysis and encodes subjective design choices even when documented; strong performance on one high-value workload can be hidden by weaker results in other categories.
Higher is better. Composite index — compare relative ranking rather than absolute values.
Weighting is set by Artificial Analysis and encodes subjective design choices even when documented.
| # | Model | Score |
|---|---|---|
| 1 | 60.0 | |
| 2 | 59.1 | |
| 3 | 58.3 | |
| 4 | 57.4 | |
| 5 | 56.9 | |
| 6 | 55.3 | |
| 7 | 54.8 | |
| 8 | 54.6 | |
| 9 | 54.5 | |
| 10 | 54.2 | |
| 11 | 53.6 | |
| 12 | 53.4 | |
| 13 | 53.0 | |
| 14 | 52.5 | |
| 15 | 51.8 | |
| 16 | 51.8 | |
| 17 | 51.6 | |
| 18 | 51.3 | |
| 19 | 50.4 | |
| 20 | 50.3 | |
| 21 | 50.1 | |
| 22 | 50.0 | |
| 23 | 49.9 | |
| 24 | 49.6 | |
| 25 | 49.2 | |
| 26 | 49.0 | |
| 27 | 48.9 | |
| 28 | 48.2 | |
| 29 | 47.2 | |
| 30 | 46.8 | |
| 31 | 46.5 | |
| 32 | 46.3 | |
| 33 | 45.3 | |
| 34 | 45.2 | |
| 35 | 45.0 | |
| 36 | 44.2 | |
| 37 | 44.0 | |
| 38 | 43.8 | |
| 39 | 43.6 | |
| 40 | 43.4 | |
| 41 | 43.3 | |
| 42 | 43.2 | |
| 43 | 42.9 | |
| 44 | 42.4 | |
| 45 | 41.9 | |
| 46 | 41.5 | |
| 47 | 38.7 | |
| 48 | 38.2 | |
| 49 | 37.1 | |
| 50 | 36.6 |