aiXamine - Adversarial Robustness: leaderboard
Metric: Service score (0-100; normalized pass rate). Source: arxiv.org. 42 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Flash (Preview) | 77.2 |
| 2 | Gemma 4 31B | 75.2 |
| 3 | GPT-5 Mini | 71.6 |
| 4 | Grok 4 Fast | 70.3 |
| 5 | Gemma 4 26B A4B | 70.1 |
| 6 | Gemma 2 27B | 69.9 |
| 7 | Qwen 3.5 27B | 69.8 |
| 8 | Qwen 3.5 122B A10B | 69.2 |
| 9 | Gemma 3 12B | 68.6 |
| 10 | Grok 4.1 Fast | 68.3 |
| 11 | Qwen 3.5 35B A3B | 68.1 |
| 12 | Llama 3.3 70B | 67.6 |
| 13 | Gemma 3 27B | 67.3 |
| 14 | DeepSeek R1 0528 | 67.2 |
| 15 | GPT-OSS-120B | 66.3 |
Interactive version: theaggregate.ai/benchmark?slug=aixamine-adversarial-robustness · How It Works · Data refreshed daily, snapshot 2026-09-19.