Trexlator
Open Data Release

Format Retention
Data Release

Our latest benchmarks on layout retention across PDF, Word, Excel, and PowerPoint translations. Real data from real documents.

View Methodology

Key Findings

Summary statistics from our comprehensive benchmark study across 12,847 test documents.

Overall Average
0%

Average layout retention across all file types and language pairs

+29.8% above industry average
Most Stable Element
Tables
0%
Most Fragile Element
Complex Charts
0%
Documents in Dataset
0

Unique documents tested across 47 language pairs

PDF
4,218
DOCX
3,941
XLSX
2,673
PPTX
2,015
Language Pairs
47

Bidirectional combinations tested

Retention by File Type

How each document format performs compared to industry average tools.

PDF

97.2%vs 68% industry avg
Layout Retention+29%
0%100%
Elements Preserved
Multi-column layoutsTables & bordersImages & graphicsHeaders/footers

Word

98.4%vs 77% industry avg
Layout Retention+21%
0%100%
Elements Preserved
Track changesCommentsStyles & formattingTable of contents

Excel

99.1%vs 54% industry avg
Layout Retention+45%
0%100%
Elements Preserved
FormulasNamed rangesConditional formattingCharts

PowerPoint

96.8%vs 69% industry avg
Layout Retention+28%
0%100%
Elements Preserved
Slide layoutsAnimationsSpeaker notesMaster slides

Element Stability Ranking

How well each layout element is preserved during translation.

1
TablesMost Stable
99.3%
2
ImagesStable
98.7%
3
Headers/FootersStable
98.2%
4
SpacingStable
97.6%
5
Page BreaksStable
96.9%
6
FontsStable
96.4%
7
Complex ChartsMost Fragile
94.1%

Methodology

We evaluated a standardized file set across multiple dimensions. Each layout element was scored on a comprehensive checklist, comparing the original document to the translated output.

  • 12,847 documents across 4 file types
  • 47 bidirectional language pairs
  • 7 layout element categories
  • Automated + manual verification
  • Anonymized enterprise documents

What's Included

Raw Data (CSV)
Element-by-element retention scores
Summary Statistics
Aggregated metrics by file type & language
Sample Documents
20 anonymized before/after examples

See It In Action

Upload your own document and see how much layout fidelity you can expect with Trexlator.

Format Retention Data Release | Translation Benchmark Data | Trexlator