F5 Labs AI Security Leaderboard - ARS: leaderboard
F5 Labs' ARS (Agentic Resistance Score, from F5 AI Red-Team) measures how well a model holds up when autonomous attack agents run multi-step, adaptive campaigns against it, rating required attacker sophistication, defensive endurance and counter-intelligence. Scale 0-100, higher is more secure. F5 publishes only its ten highest-scoring models each month, so this board is a monthly top-10 snapshot.
Metric: ARS (Agentic Resistance Score), 0-100, higher is more secure (October 2026 top-10 snapshot). Source: www.f5.com. Saturation forecast: Estimated already saturated. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 | 94.59 | #148 |
| 2 | Claude Haiku 4.5 | 93.15 | #287 |
| 3 | GPT-5.4 Mini | 91.55 | #228 |
| 4 | Claude Sonnet 4.6 | 88.23 | #94 |
| 5 | Claude Opus 4.6 | 88.18 | #67 |
| 6 | GPT-5.6 Terra | 87.46 | #40 |
| 7 | GPT-6 Luna | 86.6 | #61 |
| 8 | GPT-6 Astra | 85.21 | #6 |
| 9 | MiMo-V2.6-Pro | 84.69 | #48 |
| 10 | GPT-5 Nano | 84.22 | #439 |
No result here: #3 Claude Opus 5.5, #7 GPT-6.1 Sol, #9 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=f5-labs-ai-security-leaderboard-ars · How It Works · Data refreshed daily, snapshot 2026-11-05.