HINTBench - Fine Risk-Step Localization: leaderboard
HINTBench (2026) asks an LLM to audit long-horizon agent trajectories for intrinsic, non-attack risks such as acting beyond what the user authorized or skipping confirmations before irreversible actions. This board scores locating the risky steps (within three steps) when twelve fine-grained risk subtypes are described, ignoring the type, as F1 in percent, higher is better.
Metric: Localization F1 (%) for fine risk-step localization (twelve risk subtypes described; a predicted risky step matches a gold step within plus or minus 3 steps; type ignored) over the 536 synthetic HINTBench agent trajectories (400 risky, 136 safe, 24 steps on average, unanimously accepted by three human verifiers) audited zero-shot by a general LLM for intrinsic, non-attack risk under a five-constraint taxonomy (goal, capability, factual, procedural, state constraints), each model at its stated reasoning setting; higher is better. Source: arxiv.org. Saturation forecast: Around September 2028. 13 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Claude Sonnet 4.6 | Thinking, 4K | 60.69 | #173 |
| 2 | Claude Opus 4.6 | Thinking, 4K | 60.04 | #113 |
| 3 | DeepSeek R1 | 57.32 | #482 | |
| 4 | Qwen 3 235B A22B | 56.9 | #628 | |
| 5 | GLM-5 | 56.3 | #269 | |
| 6 | ERNIE 5.0 | 56.16 | #479 | |
| 7 | Kimi K2.5 | 56.15 | #266 | |
| 8 | GPT-5.5 | Low | 54.17 | #54 |
| 9 | DeepSeek V3.2 | 54.07 | #419 | |
| 10 | MiniMax-M2.5 | 53.82 | #600 | |
| 11 | Qwen 3 32B | 51.31 | #972 | |
| 12 | Qwen 3 14B | 50.65 | #1269 | |
| 13 | Qwen 3 8B | 29.71 | #1873 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=hintbench-fine-risk-step-localization · How It Works · Data refreshed daily, snapshot 2026-11-21.