DTap-Bench (Direct Attacks) - Windows: leaderboard
Metric: Attack success rate (%; 140 direct red-teaming Windows desktop tasks: the user is the attacker, with prompt, tool, skill and environment injections; an attack succeeds when a rule-based check of the final environment state finds the malicious goal fully achieved; fixed attack sequences that DTap-Red optimized on a surrogate agent; mean of its risk-category rates). Source: arxiv.org. Saturation forecast: Around 2030. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-OSS-120B | 33.3 |
| 2 | Gemini 3 Pro | 38.8 |
| 3 | GPT-5.2 | 44.6 |
| 4 | GPT-5.4 | 47.9 |
Interactive version: theaggregate.ai/benchmark?slug=dtap-bench-direct-attacks-windows · How It Works · Data refreshed daily, snapshot 2026-09-26.