Turkish MMLU - KPSS: leaderboard
100 multiple-choice questions from KPSS, the Turkish public-service personnel selection exam, one of 62 sections of the Turkish MMLU leaderboard, whose models run mostly as quantized Ollama builds. Percent correct, higher is better.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 59 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 72 | #357 |
| 2 | Claude 3.5 Sonnet (20240620) | 65 | #308 |
| 3 | Gemini 1.5 Pro | 62 | #454 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=turkish-mmlu-kpss · How It Works · Data refreshed daily, snapshot 2026-11-05.