THOR Finding Triage - Balanced OTS: leaderboard
Nextron's THOR Finding Triage Benchmark asks models to triage 189 expert-labelled findings from THOR forensic scanner reports. Balanced OTS is the class-balanced operational triage score in percent; higher is better. Naive always-one-class baselines are excluded.
Metric: Class-balanced operational triage score (%). Source: nextron-labs.github.io. Saturation forecast: Around November 2027. 89 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.7 Flash | 72.5 | #24 |
| 2 | Gemini 3.1 Flash Lite | 70.8 | #200 |
| 3 | GLM-5.1 | 69.6 | #109 |
| 4 | Claude Opus 4.5 | 69 | #87 |
| 5 | Gemini 3.5 Flash | 68.8 | #66 |
| 6 | Gemma 4 31B | 68.5 | #212 |
| 7 | Kimi K2.7 Code | 67.7 | #95 |
| 8 | Gemini 3.6 Flash | 66.9 | #57 |
| 9 | DeepSeek V4 Pro (0813) | 66.9 | #58 |
| 10 | Gemini 3.1 Pro (Preview) | 66.3 | #64 |
| 11 | DeepSeek V4.1 Flash (1017) | 66.1 | #27 |
| 12 | GLM-5 | 65.5 | #149 |
| 13 | GLM-5.3 | 65.2 | #44 |
| 14 | Claude Opus 4.6 | 64.8 | #67 |
| 15 | Gemini 3.8 Flash | 64.72 | #30 |
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-balanced-ots · How It Works · Data refreshed daily, snapshot 2026-11-05.