Skip to content

Benchmark

ArmBench-LLM Armenian squad (bleu, 0-shot)

Higher scores are better.

Models measured
27
Scores recorded
28
Sources
1
IQ weight
0.03

Score distribution

3.9
Mean
4.2
Median
2
Std dev
1.4
10th percentile
5.7
90th percentile
0.525
Min
0.5258.8
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.03 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 docs8.8
2Qwen3.5 9B· ArmBench-LLM Armenian data docs7.5
3Grok 4.1 Fast· ArmBench-LLM Armenian data docs6.3
4Qwen3.5 27B· ArmBench-LLM Armenian data docs5.7
5Gemma 3 12B IT· ArmBench-LLM Armenian data docs5.6
6GPT-5.4 mini· ArmBench-LLM Armenian data docs5.4
7GLM 4.7 Flash· ArmBench-LLM Armenian data docs5.2
8Gemma 4 31B IT· ArmBench-LLM Armenian data docs5.1
9Gemma 3 4B IT· ArmBench-LLM Armenian data docs5.1
10Gemma 3 27B IT· ArmBench-LLM Armenian data docs4.8
11Gemini 2.5 Pro· ArmBench-LLM Armenian data docs4.5
12Gemini 2.5 Flash· ArmBench-LLM Armenian data docs4.3
13GLM 5· ArmBench-LLM Armenian data docs4.3
14Gemini 3 Pro Preview· ArmBench-LLM Armenian data docs4.2
15GPT 5.4 Nano· ArmBench-LLM Armenian data docs4.1
16Gemini 3.1 Pro Preview· ArmBench-LLM Armenian data docs3.6
17Qwen3 4B Instruct 2507· ArmBench-LLM Armenian data docs3.6
18Gemini 3 Flash· ArmBench-LLM Armenian data docs3.4
19Claude Sonnet 3.7· ArmBench-LLM Armenian data docs2.7
20Claude Sonnet 4· ArmBench-LLM Armenian data docs2.5
21GPT-5.2 Prosetting not named by the source· ArmBench-LLM Armenian data docs2.4
22GPT-5.2 Prosetting not named by the source· ArmBench-LLM Armenian data docs2.4
23Mistral Large 3· ArmBench-LLM Armenian data docs1.7
24Qwen3 4B Base· ArmBench-LLM Armenian data docs1.4
25Mistral Small 3.2 24B Instruct· ArmBench-LLM Armenian data docs1.4
26Grok 4.20· ArmBench-LLM Armenian data docs1.4
27Ministral 14B· ArmBench-LLM Armenian data docs0.537
28Qwen3.5 35B A3B· ArmBench-LLM Armenian data docs0.525

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