Swiss-Bench 003 - Adversarial Security: leaderboard
Metric: D8 adversarial security composite (%): 0.6 x PII-Scope score + 0.4 x System Prompt Leakage score, over 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.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-OSS-120B | 60.7 |
| 2 | GPT-4o | 55.1 |
| 3 | Qwen 3.5 Plus | 50.6 |
| 4 | Claude Sonnet 4 | 44.2 |
| 5 | MiniMax-M2.5 | 44 |
| 6 | GLM-5 | 43.2 |
| 7 | MiMo-V2-Flash | 37.4 |
| 8 | Gemini 2.5 Flash | 27.9 |
| 9 | Mistral Large 3 | 23 |
| 10 | DeepSeek V3.2 | 20.1 |
Interactive version: theaggregate.ai/benchmark?slug=swiss-bench-003-adversarial-security · How It Works · Data refreshed daily, snapshot 2026-10-07.