Phare - Jailbreak Resistance: leaderboard
Giskard’s Phare multilingual safety probe (English, French, Spanish), Jailbreak Resistance module: whether policy holds under jailbreak framing. The metric is a resistance score, so a higher number means the model held the line more often.
Metric: Score (%). Source: phare.giskard.ai. Saturation forecast: Around January 2027. 68 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 5 | 86.84 | #57 |
| 2 | Claude Opus 4.1 | 81.35 | #140 |
| 3 | Claude Opus 4.5 | 79.83 | #83 |
| 4 | Claude Haiku 4.5 | 78.51 | #280 |
| 5 | Claude Opus 4.6 | 75.24 | #63 |
| 6 | Claude Sonnet 4.5 | 75.23 | #142 |
| 7 | Qwen 3.7 Max | 73.16 | #75 |
| 8 | GPT-5.2 | 71.61 | #110 |
| 9 | Kimi K2.6 | 71.4 | #104 |
| 10 | Qwen 3.7 Plus | 70.58 | #92 |
| 11 | GPT-5 Nano | 69.64 | #425 |
| 12 | Claude Sonnet 4.6 | 69.39 | #90 |
| 13 | GPT-5 | 68.68 | #98 |
| 14 | GPT-5 Mini | 68.6 | #186 |
| 15 | GPT-OSS-120B | 65.9 | #341 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=phare-jailbreak-resistance · How It Works · Data refreshed daily, snapshot 2026-10-20.