Skip to content

Benchmark

KazBench Kazakh grammar_morphology (accuracy, dev)

Higher scores are better.

Models measured
6
Scores recorded
6
Sources
1
IQ weight
0.45

Score distribution

0.868
Mean
0.917
Median
0.082
Std dev
0.771
10th percentile
0.979
90th percentile
0.75
Min
0.750.979
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.45 for chat models, measured over 6 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Haiku 3.5dev, version: 0.1.0· KazBench Kazakh data docs0.979
2Gemini 2.5 Flashdev, version: 0.1.0· KazBench Kazakh data docs0.917
3GPT-4o minidev, version: 0.1.0· KazBench Kazakh data docs0.917
4Llama 4 Scoutdev, version: 0.1.0· KazBench Kazakh data docs0.875
5Llama 3.1 8B Instantdev, version: 0.1.0· KazBench Kazakh data docs0.771
6Qwen2.5 72B Instructdev, version: 0.1.0· KazBench Kazakh data docs0.75

Data credits: KazBench by Yersultan Akhmer. Provisional DEV numeric results; code MIT and task-data CC BY 4.0 do not establish a separate score licence.