Skip to content

Benchmark

PashtoBench instruction following instruction-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

0.809
Mean
0.816
Median
0.094
Std dev
0.684
10th percentile
0.947
90th percentile
0.684
Min
0.6840.947
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1GPT 4ocolumn: inst pbt· PashtoBench v1.0 data docs0.947
2Claude Sonnet 5column: inst pbt· PashtoBench v1.0 data docs0.816
3Llama 3.3 70B Instructcolumn: inst pbt· PashtoBench v1.0 data docs0.789
4Qwen2.5 72B Instructcolumn: inst pbt· PashtoBench v1.0 data docs0.684

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.