Enkrypt AI - Bias Risk: leaderboard
Enkrypt AI’s LLM safety leaderboard: measured bias in answers to demographically loaded prompts. It is reported as a risk, so lower is better here and the leader is the least risky model rather than the most capable one.
Metric: Risk Score. Source: leaderboard.enkryptai.com. Saturation forecast: Around 2031. 300 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5 | 0 | #16 |
| 2 | Mistral-NeMo-Minitron-8B-Instruct | 7.49 | #907 |
| 3 | Mistral-7B-v0.1 | 13.44 | #1326 |
| 4 | SmolLM-360M-Instruct | 23.77 | #1597 |
| 5 | Claude 3 Opus (20240229) | 25.58 | #412 |
| 6 | Claude Opus 5.5 | 25.58 | #3 |
| 7 | Grok 4.7 | 26.61 | #25 |
| 8 | Claude Sonnet 5 | 29.46 | #57 |
| 9 | Grok 4.6 | 35.14 | #26 |
| 10 | Claude 3.5 Sonnet (20241022) | 35.4 | #291 |
| 11 | GPT-5 | 40.05 | #98 |
| 12 | Claude Opus 5 | 40.83 | #12 |
| 13 | Claude 3.5 Haiku (20241022) | 43.41 | #588 |
| 14 | Qwen 2.5 72B Instruct Turbo | 44.19 | #527 |
| 15 | MiniMax-M2.7 | 45.48 | #258 |
No result here: #6 GPT-6 Astra, #7 GPT-6.1 Sol, #8 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=enkrypt-ai-bias-risk · How It Works · Data refreshed daily, snapshot 2026-10-20.