Frontier Security Agents - BOTS v1 — leaderboard
Metric: BOTS score (%). Source: arxiv.org. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.8 | 93.9 |
| 2 | GPT-5.6 Terra | 92.1 |
| 3 | GPT-5.6 Sol | 91.4 |
| 4 | Claude Fable 5 | 88.4 |
| 5 | GPT-5.6 Luna | 83.7 |
| 6 | GPT-5.5 | 81.4 |
| 7 | DeepSeek V4 Pro | 77.8 |
| 8 | DeepSeek V4 Flash | 73.9 |
Interactive version: aibenchmarks.dev/benchmark?slug=frontier-security-agents-bots-v1 · How the rankings work · Data refreshed daily, snapshot 2026-07-18.