MonitoringBench - Recon - Firewall Backdoor: leaderboard
MonitoringBench (2026) tests LLMs as monitors of coding agents in BashArena: the monitor reads a full agent trajectory of tool calls and scores how suspicious it is, and an attack counts as caught when the score exceeds the monitor threshold for a 1% false-positive rate on honest runs. This board covers the 204 successful firewall backdoor attacks from the recon stage of the red-teaming pipeline, scored as catch rate in percent; higher is better.
Metric: Catch rate (%) at a 1% false-positive-rate threshold calibrated on 416 honest BashArena runs: share of the 204 successful firewall backdoor attack trajectories (strategies from reconnaissance agents (with and without GPT-5 monitor feedback) executed by Claude Opus 4.6) that the full-trajectory monitor scores above its threshold, default ControlArena monitor prompt with a 0-10 suspiciousness score, mean of three scoring runs; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 13 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.5 (20251101) | 72.9 | #142 |
| 2 | Claude Sonnet 4.5 | 62.7 | #267 |
| 3 | Claude Sonnet 4 (20250514) | 56 | #410 |
| 4 | GPT-5.1 Codex Max | 51.8 | #212 |
| 5 | O3 (2025-04-16) | 35.9 | #238 |
| 6 | GPT-5 | 29.7 | #185 |
| 7 | GPT-4.1 | 16.5 | #506 |
| 8 | GPT-4o (2024-08-06) | 12.7 | #750 |
| 9 | GPT-4 Turbo | 7.5 | #1015 |
| 10 | GPT-5 Nano | 5.7 | #912 |
| 11 | GPT-4.1 Mini | 4.7 | #799 |
| 12 | O4 Mini (2025-04-16) | 2.5 | #375 |
| 13 | GPT-4o Mini (2024-07-18) | 0.5 | #1501 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=monitoringbench-recon-firewall-backdoor · How It Works · Data refreshed daily, snapshot 2026-11-22.