Skip to content

Benchmark

PashtoBench instruction following prompt-level accuracy (English)

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

0.704
Mean
0.727
Median
0.039
Std dev
0.636
10th percentile
0.727
90th percentile
0.636
Min
0.6360.727
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1Claude Sonnet 5column: prompt en· PashtoBench v1.0 data docs0.727
2GPT 4ocolumn: prompt en· PashtoBench v1.0 data docs0.727
3Llama 3.3 70B Instructcolumn: prompt en· PashtoBench v1.0 data docs0.727
4Qwen2.5 72B Instructcolumn: prompt en· PashtoBench v1.0 data docs0.636

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.