Skip to content

Benchmark

ArmBench-LLM Armenian exam_math (armenian_exam_score, 0-shot)

Higher scores are better.

Models measured
27
Scores recorded
28
Sources
1
IQ weight
1.12

Score distribution

13.8
Mean
14.8
Median
4.3
Std dev
10.5
10th percentile
18.5
90th percentile
1.8
Min
1.818.8
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.12 for chat models, measured over 27 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Grok 4 Fast· ArmBench-LLM Armenian data docs18.8
2Grok 4.1 Fast· ArmBench-LLM Armenian data docs18.8
3GPT-5.2 Prosetting not named by the source· ArmBench-LLM Armenian data docs18.5
4GPT-5.2 Prosetting not named by the source· ArmBench-LLM Armenian data docs18.5
5Qwen3.5 27B· ArmBench-LLM Armenian data docs18
6Gemma 4 31B IT· ArmBench-LLM Armenian data docs17.3
7Gemini 3 Flash· ArmBench-LLM Armenian data docs16.8
8Gemini 3 Pro Preview· ArmBench-LLM Armenian data docs16.3
9Gemma 3 27B IT· ArmBench-LLM Armenian data docs16.3
10Mistral Large 3· ArmBench-LLM Armenian data docs16
11Claude Sonnet 3.7· ArmBench-LLM Armenian data docs15.3
12Gemini 2.5 Pro· ArmBench-LLM Armenian data docs15.3
13Claude Sonnet 4· ArmBench-LLM Armenian data docs15
14Gemma 3 12B IT· ArmBench-LLM Armenian data docs14.8
15Gemini 2.5 Flash· ArmBench-LLM Armenian data docs14.5
16Gemini 3.1 Pro Preview· ArmBench-LLM Armenian data docs14.5
17Ministral 14B· ArmBench-LLM Armenian data docs14.3
18GLM 4.7 Flash· ArmBench-LLM Armenian data docs13.5
19Mistral Small 3.2 24B Instruct· ArmBench-LLM Armenian data docs13.5
20GLM 5· ArmBench-LLM Armenian data docs13.3
21Qwen3.5 35B A3B· ArmBench-LLM Armenian data docs13
22GPT-5.4 mini· ArmBench-LLM Armenian data docs11.5
23Qwen3 4B Instruct 2507· ArmBench-LLM Armenian data docs11.5
24Gemma 3 4B IT· ArmBench-LLM Armenian data docs11
25GPT 5.4 Nano· ArmBench-LLM Armenian data docs10.5
26Qwen3.5 9B· ArmBench-LLM Armenian data docs7.8
27Grok 4.20· ArmBench-LLM Armenian data docs1.8
28Qwen3 4B Base· ArmBench-LLM Armenian data docs1.8

Data credits: ArmBench-LLM by Metric-AI. Public numeric primary task results; no separate results licence asserted.