Kaggle MultiLoKo - Indonesian: leaderboard

Metric: Exact Match (%). Source: www.kaggle.com. 48 models tracked.

Top models

#ModelScore
1Gemini 3.1 Pro (Preview)62.8
2Gemini 3.5 Flash55.2
3Gemini 3 Flash (Preview)51.6
4Gemini 2.5 Pro (Preview 06-05)49.2
5GPT-542
6O340.4
7O138.4
8Gemini 3.1 Flash Lite (Preview)38
9Gemini 2.5 Flash (Preview 05-20)36
10GPT-5.432.4
11GPT-4o (2024-08-06)31.6
12Gemini 2.0 Flash30.4
13GPT-4.130
14Gemini 1.5 Pro (002)27.6
15Claude 3.5 Sonnet (20241022)26.4

Interactive version: theaggregate.ai/benchmark?slug=kaggle-multiloko-indonesian · How It Works · Data refreshed daily, snapshot 2026-09-19.