Skip to content

Benchmark

PashtoBench factual QA token F1 (Pashto)

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

0.078
Mean
0.115
Median
0.068
Std dev
0.007
10th percentile
0.17
90th percentile
0.007
Min
0.0070.17
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1Llama 3.3 70B Instructcolumn: F1 pbt· PashtoBench v1.0 data docs0.17
2GPT 4ocolumn: F1 pbt· PashtoBench v1.0 data docs0.115
3Claude Sonnet 5column: F1 pbt· PashtoBench v1.0 data docs0.018
4Qwen2.5 72B Instructcolumn: F1 pbt· PashtoBench v1.0 data docs0.007

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.