Kaggle Global-MMLU-Lite - Japanese: leaderboard
Metric: Accuracy (%). Source: www.kaggle.com. 37 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 5 | 96 |
| 2 | Gemini 3.5 Flash | 95.75 |
| 3 | GPT-6 | 95.5 |
| 4 | Claude Opus 4.8 | 94.75 |
| 5 | GPT-5.6 Sol | 94.75 |
| 6 | Grok 4.5 | 94.75 |
| 7 | Gemini 3.6 Flash | 94.75 |
| 8 | Gemini 3.8 Flash | 94.75 |
| 9 | Gemini 3.7 Flash | 94.5 |
| 10 | Gemini 3 Pro (Preview) | 94 |
| 11 | Claude Opus 4.1 | 94 |
| 12 | Gemini 2.5 Pro | 93.75 |
| 13 | GPT-5.6 Terra | 93.75 |
| 14 | Gemini 2.5 Flash | 91.25 |
| 15 | Gemini 2.5 Flash (Preview 05-20) | 91.25 |
Interactive version: theaggregate.ai/benchmark?slug=kaggle-global-mmlu-lite-japanese · How It Works · Data refreshed daily, snapshot 2026-09-19.