THOR Finding Triage - False Review Load: leaderboard
Nextron's THOR Finding Triage Benchmark asks models to triage 189 expert-labelled findings from THOR forensic scanner reports. False Review Load is the percentage of false-positive findings the model fails to suppress, leaving review work for analysts; lower is better.
Metric: False positives not suppressed (%, lower is better). Source: nextron-labs.github.io. Saturation forecast: Not forecast. 89 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.6 Max | 22.7 | #97 |
| 2 | Qwen 3.6 35B A3B | 23.7 | #238 |
| 3 | Qwen 3.6 27B | 24.7 | #182 |
| 4 | Qwen 3.6 Plus | 26.8 | #124 |
| 5 | Gemini 3.6 Flash | 27.8 | #57 |
| 6 | Qwen 3.7 Plus | 28.9 | #98 |
| 7 | Gemini 3.7 Flash | 28.9 | #24 |
| 8 | Qwen 3.6 Flash | 29.9 | #221 |
| 9 | Kimi K2.6 | 30.9 | #113 |
| 10 | Gemini 3.5 Flash | 30.9 | #66 |
| 11 | MiMo-V2-Pro | 30.9 | #177 |
| 12 | Qwen 3.7 Max | 32 | #85 |
| 13 | Qwen 3.5 9B | 33 | #391 |
| 14 | Claude Opus 4.8 | 33 | #39 |
| 15 | Gemini 3.1 Flash Lite | 33 | #200 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=thor-finding-triage-false-review-load · How It Works · Data refreshed daily, snapshot 2026-11-05.