Kaggle MultiLoKo - Cantonese: leaderboard

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

Top models

#ModelScore
1Gemini 3.1 Pro (Preview)55.33
2Gemini 3.5 Flash46.18
3Gemini 3 Flash (Preview)44.8
4Gemma 4 26B A4B44.4
5Gemini 2.5 Pro (Preview 06-05)37.2
6Gemini 3.1 Flash Lite (Preview)34
7GPT-530.4
8DeepSeek R1 052830.4
9O328.8
10O128
11Gemini 2.5 Flash (Preview 05-20)28
12DeepSeek V324.4
13Gemini 2.0 Flash23.2
14GPT-5.422.8
15Claude Opus 422.8

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