Skip to content

Benchmark

Phare Jailbreak Resistance

Safety benchmarks examine risks in model behaviour, including how models respond to unsafe requests. Scores are percentages.

Models measured
67
Scores recorded
67
Sources
1
IQ weight
0.97

Score distribution

52.5%
Mean
49%
Median
15.6%
Std dev
34.4%
10th percentile
73.2%
90th percentile
24.9%
Min
24.9%86.8%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.97 for chat models, measured over 67 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

Showing 1–50 of 67 results.

Data credits: Phare (Giskard), CC BY 4.0 (licence)