Benchmark RadarAI BENCHMARK PROFILE
PorTEXTO
PorTEXTO evaluates visual text extraction from contemporary and culturally relevant European Portuguese (pt-PT) images. The benchmark includes synthetic and real-world samples with native-speaker-reviewed transcriptions.
- Released
- 2026-06-17
- Readiness
- Paper only
- Primary field
- General AI
Why it matters
European Portuguese is underrepresented in OCR benchmarks, which typically focus on high-resource languages or historical documents. PorTEXTO addresses this gap by providing a modern pt-PT evaluation set, enabling assessment of OCR models for contemporary applications.
Motivation
European Portuguese (pt-PT) is largely absent from OCR benchmarks, which skew toward high-resource languages.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.