Swiss-Bench 003 - PII Extraction Defense: leaderboard
Defense against Swiss personal-data extraction on Swiss-Bench 003 (2026): over 271 social-engineering prompts in four languages, a Qwen3-235B judge scores a refusal 1, a plausible but fake identifier 0.5 and generated Swiss personal data such as AHV numbers or IBANs 0. Mean score in percent; higher is better.
Metric: Swiss PII-Scope score (%): mean of 1.0 for a refusal, 0.5 for a plausible-format but not real identifier and 0.0 for generated Swiss personal data (AHV/AVS numbers, Swiss IBANs, cantonal identifiers) over 271 social-engineering prompts, part of the Swiss-Bench 003 (SBP-003) adversarial security (D8) items in German, French, Italian and English, zero-shot at provider defaults, judged by Qwen3-235B with tiered rubrics; higher is better. Source: arxiv.org. Saturation forecast: Not forecast. 10 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-OSS-120B | 42.4 | #712 |
| 2 | Qwen 3.5 Plus | 40.4 | #250 |
| 3 | GPT-4o | 38.7 | #754 |
| 4 | Claude Sonnet 4 | 38.6 | #414 |
| 5 | MiniMax-M2.5 | 33.2 | #600 |
| 6 | Gemini 2.5 Flash | 27.9 | #466 |
| 7 | GLM-5 | 24.9 | #276 |
| 8 | MiMo-V2-Flash | 23.6 | #691 |
| 9 | Mistral Large 3 | 21.8 | #835 |
| 10 | DeepSeek V3.2 | 14.2 | #425 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=swiss-bench-003-pii-extraction-defense · How It Works · Data refreshed daily, snapshot 2026-11-22.