SkillTrustBench v1.1 - AIG - Malicious Recall: leaderboard
SkillTrustBench v1.1 Malicious Recall for nine identity-bound model runs inside Tencent AI-Infra-Guard. The fixed public subset has 556 agent-skill cases: 316 malicious, 74 suspicious and 166 normal. Native project verdicts are preferred; empty reports count normal and missing verdicts may use severity fallback. The scored content manifest is pinned; exact framework revision, reasoning effort and budget are unreported. The ambiguous DeepSeek 0910/generic-route row is held separately.
Metric: Malicious Recall (0–1). Source: huggingface.co. Saturation forecast: Rough model projection: around 2026. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | MiniMax-M3 (SkillTrustBench Run: AIG v1.1; 2026-10-25; effort and budget unreported) | 0.96 | |
| 2 | Kimi K3 (SkillTrustBench Run: AIG v1.1; 2026-10-25; effort and budget unreported) | 0.94 | |
| 3 | DeepSeek V4.1 Flash (1017) (SkillTrustBench Run: AIG v1.1; 2026-10-25; effort and budget unreported) | 0.94 | |
| 4 | GLM-5.3 (SkillTrustBench Run: AIG v1.1; 2026-10-25; effort and budget unreported) | 0.93 | |
| 5 | Hy4 preview (SkillTrustBench Run: AIG v1.1; 2026-10-25; effort and budget unreported) | 0.91 | |
| 6 | Claude Haiku 5.5 (SkillTrustBench Run: AIG v1.1; 2026-10-25; effort and budget unreported) | 0.86 | |
| 7 | MiMo-V2.6-Flash (SkillTrustBench Run: AIG v1.1; 2026-10-25; effort and budget unreported) | 0.81 | |
| 8 | GPT-6 Luna (SkillTrustBench Run: AIG v1.1; 2026-10-25; effort and budget unreported) | 0.7 | |
| 9 | MiniMax-M3.1-Flash (SkillTrustBench Run: AIG v1.1; 2026-10-25; effort and budget unreported) | 0.64 |
Interactive version: theaggregate.ai/benchmark?slug=skilltrustbench-v1-1-aig-malicious-recall · How It Works · Data refreshed daily, snapshot 2026-11-05.