Skip to content

Benchmark

PashtoBench translation English to Pashto BLEU

Multilingual benchmarks compare model performance across languages. Higher scores are better.

Models measured
4
Scores recorded
4
Sources
1
IQ weight
No weight published

Score distribution

28.2
Mean
50
Median
26.4
Std dev
0.5
10th percentile
58.8
90th percentile
0.5
Min
0.558.8
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1GPT 4odirection: en-pbt· PashtoBench v1.0 data docs58.8
2Claude Sonnet 5direction: en-pbt· PashtoBench v1.0 data docs50
3Qwen2.5 72B Instructdirection: en-pbt· PashtoBench v1.0 data docs3.5
4Llama 3.3 70B Instructdirection: en-pbt· PashtoBench v1.0 data docs0.5

Data credits: PashtoBench v1.0, Jawad Noori. Public results/scores numbers; code MIT and data/ CC BY 4.0 grants do not name results/, so no results licence is asserted.