TriBench-Ko - Statutory Misapplication: leaderboard

The statutory misapplication risk dimension of TriBench-Ko (2026), a Korean benchmark of attorney-reviewed yes/no items built from court decisions, where false statements carry provisions applied where they do not fit the facts. The score is macro-F1 on a 0-100 scale over the atomic binary judgments, zero-shot at temperature 0; higher is better.

Metric: Macro-F1 (0-100) on the TriBench-Ko Statutory Misapplication risk items (provisions applied where they do not fit the facts): yes/no verification of statements about Korean court decisions, computed over the atomic binary judgments, zero-shot, temperature 0, at most 64 output tokens; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 13 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-5.482.2#135
2GPT-5.4 Mini72.9#406
3Qwen 3.5 9B69.7#828
4Phi-469.3#1931
5GPT-4o67.3#751
6Qwen 3 8B66.3#1873
7Ministral-3-8B-Instruct-251264.3#1907
8Gemma 3 12B (IT)63.7#1887
9Midm-2.0-Base-Instruct63.7#1988
10EXAONE 3.5 7.8B Instruct43.2#3410
11kanana-1.5-8B-instruct-250539.6#3315
12Llama 3.1 8B Instruct34.3#4600
13A.X-3.1-Light32.6

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=tribench-ko-statutory-misapplication · How It Works · Data refreshed daily, snapshot 2026-11-21.