Skip to content

Benchmark

PashtoBench instruction following prompt-level accuracy (Pashto)

Instruction following benchmarks measure how closely models follow instructions. Higher scores are better.

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

Score distribution

0.758
Mean
0.767
Median
0.119
Std dev
0.6
10th percentile
0.933
90th percentile
0.6
Min
0.60.933
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1GPT 4ocolumn: prompt pbt· PashtoBench v1.0 data docs0.933
2Claude Sonnet 5column: prompt pbt· PashtoBench v1.0 data docs0.767
3Llama 3.3 70B Instructcolumn: prompt pbt· PashtoBench v1.0 data docs0.733
4Qwen2.5 72B Instructcolumn: prompt pbt· PashtoBench v1.0 data docs0.6

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.