| # | Model | Overall Score | Language | Safety | Healthcare | Public Sector |
|---|---|---|---|---|---|---|
| 1 | gpt-5-mini | 92.7 ±5.3 | 87.5 | 90.6 | 92.5 | 100.0 |
| 2 | NbAiLab/borealis-27b | 65.7 ±11.0 | 60.3 | 70.3 | 78.5 | 53.5 |
| 3 | NbAiLab/borealis-open-27b | 63.2 ±11.5 | 57.2 | 66.2 | 77.8 | 51.5 |
| 4 | NbAiLab/borealis-27b-instruct-preview | 61.2 ±9.1 | 59.0 | 68.8 | 67.8 | 49.2 |
| 5 | NbAiLab/borealis-open-12b | 60.5 ±9.4 | 54.4 | 65.3 | 71.2 | 51.0 |
| 6 | NbAiLab/borealis-12b | 59.8 ±9.4 | 54.1 | 62.5 | 71.8 | 50.8 |
| 7 | google/gemma-3-27b-it | 59.2 ±4.8 | 62.8 | 63.8 | 55.8 | 54.5 |
| 8 | norallm/normistral-11b-thinking | 57.9 ±5.4 | 61.6 | 50.9 | 65.2 | 53.2 |
| 9 | google/gemma-3-12b-it | 57.4 ±6.7 | 55.0 | 62.8 | 62.8 | 49.0 |
| 10 | NbAiLab/borealis-12b-instruct-preview | 57.1 ±7.9 | 54.7 | 57.8 | 67.5 | 48.5 |
| 11 | NbAiLab/borealis-4b-instruct-preview | 39.7 ±6.7 | 42.5 | 33.1 | 47.8 | 35.5 |
| 12 | NbAiLab/borealis-4b | 39.3 ±4.9 | 45.9 | 38.1 | 39.0 | 34.2 |
| 13 | NbAiLab/borealis-open-4b | 37.5 ±7.8 | 41.2 | 37.8 | 44.2 | 26.5 |
| 14 | google/gemma-3-4b-it | 25.5 ±10.0 | 39.7 | 16.9 | 24.5 | 20.8 |
| 15 | NbAiLab/borealis-open-1b | 8.7 ±9.1 | 22.2 | 3.1 | 6.0 | 3.5 |
| 16 | NbAiLab/borealis-1b | 8.3 ±9.3 | 22.2 | 3.4 | 4.5 | 3.0 |
| 17 | NbAiLab/borealis-1b-instruct-preview | 8.1 ±7.5 | 19.1 | 5.6 | 5.2 | 2.5 |
| 18 | google/gemma-3-1b-it | 6.4 ±9.3 | 20.3 | 0.9 | 3.2 | 1.2 |
| 19 | google/gemma-3-270m-it | 6.3 ±7.5 | 17.5 | 1.9 | 2.8 | 3.2 |
| 20 | NbAiLab/borealis-270m-instruct-preview | 4.7 ±7.1 | 15.3 | 1.9 | 0.5 | 1.0 |
| 21 | NbAiLab/borealis-open-270m | 4.3 ±6.3 | 13.8 | 1.2 | 1.2 | 1.0 |
| 22 | NbAiLab/borealis-270m | 3.6 ±4.5 | 10.3 | 1.2 | 1.8 | 1.0 |
Per-Pack Scores
| Pack | Score | Severity Breakdown (mean per run) |
|---|---|---|
| Language | 61.6 ±8.0 | High: 2.0, Pass: 2.6, Medium: 1.7, Low: 1.3, Critical: 0.3, Error: 0.2 |
| Safety | 50.9 ±10.6 | High: 2.8, Critical: 1.2, Pass: 2.6, Medium: 1.1, Low: 0.3 |
| Healthcare | 65.2 ±10.2 | Critical: 1.2, Pass: 5.4, High: 2.3, Medium: 1.1 |
| Public Sector | 53.2 ±7.2 | Medium: 1.1, High: 4.4, Pass: 3.6, Low: 0.1, Critical: 0.8 |
Severity Distribution (overall, mean per run)
| Critical | 3.5 ±2.0 (n=35) | |
| High | 11.5 ±2.5 (n=115) | |
| Medium | 5.0 ±1.1 (n=50) | |
| Low | 1.7 ±0.9 (n=17) | |
| Pass | 14.2 ±2.2 (n=142) |