We created a sample DOCX containing a range of document structures, processed the same file in Microsoft Word and Thinkfree Office, and compared the resulting DOCX and PDF files.
SINGLE TEST DOCUMENT · DOCX + PDF · AUGUST 2026
Visual comparison is only the first check. We also compared the structure of the resulting DOCX files and the layout of the exported PDFs.
Microsoft Word
Thinkfree Document Editor SDK
same page, same zoom
The same source DOCX was processed in Microsoft Word and Thinkfree Office. We then analyzed the resulting document packages and compared the structures present in this sample.
Results below apply to this sample document.
19 / 19
OOXML parts preserved
All 19 OOXML parts checked in this sample remained present after the round trip.
16 / 16
Construct counts unchanged
Counts for merged cells, cell borders, section breaks, page-number fields, and body text remained unchanged in this sample.
1
Difference found
One difference in the round trip.
| Construct | OOXML | Input | Output | Result |
|---|---|---|---|---|
| Footnote | word/footnotes.xml | 1 | 1 | Identical |
| Embedded chart | word/charts/chart1.xml | 3 series | 3 series | 27 / 27 values |
| Row merge | w:vMerge | 4 | 4 | Identical |
| Column merge | w:gridSpan | 1 | 1 | Identical |
| PAGE field | w:fldChar + w:instrText | 2 | 2 | Still a field |
| Headers / footers | header1–3, footer1–3 | 3 + 3 | 3 + 3 | Identical |
| Cell borders / shading | w:tcBorders, w:shd | 66 / 6 | 66 / 6 | Identical |
| Fixed table layout | w:tblLayout, w:tcW | 1 / 33 | 1 / 33 | Identical |
| Section breaks | w:sectPr | 2 | 2 | Identical |
| Body text | w:t runs | 41 | 41 | Byte-identical |
| List definitions | w:abstractNum, w:num | 2 / 3 | 1 / 1 | Consolidated |
WHAT CHANGED
List definitions were consolidated.
The input contained two abstract list definitions and three list instances. The resulting document contained one abstract definition and one list instance. The visible bullet formatting remained unchanged in this sample.
We also exported the same sample from Microsoft Word and Thinkfree Document SDK and compared the resulting PDFs.
49 / 50
text runs within 0.5pt
In this sample.
0.09pt
vertical offset standard deviation
Across matching text runs in this sample.
A small number of export differences were also observed in this sample.
TEST METHODOLOGY
A synthetic DOCX sample containing a range of document structures was created for this comparison. The same source file was opened and processed in Microsoft Word and Thinkfree Document SDK. The resulting DOCX files and PDFs were then analyzed and compared.
The sample and resulting files were analyzed with Claude. These results describe this test document only and should not be read as a general fidelity benchmark across all DOCX files or features.
What this sample doesn’t cover
Hyperlinks, raster images, tracked changes, comments, text boxes, CJK line breaking, and more complex ordered-list behavior were outside the scope of this sample.
SAMPLE DOCUMENT
Download the source and resulting files used in this comparison and inspect them directly.
