Greek NLU Benchmark - NLU Macro: leaderboard
Greek NLU Benchmark v1.0 by KIEFER TEK: native-verified Greek items over 17 language-understanding tasks (acceptability, cloze, coreference, idioms, NLI, paraphrase, sentiment, word sense and others), weighted toward morphology. Macro-average of the task scores, 0-100, higher is better; a small board of seven models.
Metric: Macro-average of 17 task scores (0-100). Source: huggingface.co. Saturation forecast: Around February 2027. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GLM-5.2 FP8 | 68.47 | #199 |
| 2 | Qwen 3.5 397B A17B | 66.77 | #154 |
| 3 | MiniMax-M3 | 65.51 | #151 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=greek-nlu-benchmark-nlu-macro · How It Works · Data refreshed daily, snapshot 2026-11-05.