The results of the shared task are summarized in the overview papers.

The raw data is available here:

• News: Submissions, references, automatic scores and human evaluations
• Similar languages: (coming soon)
• Biomedical: Test sets, gold sets, alignments, submissions
• European low res multilingual: results, validation set, test set
• Large-scale multilingual: (coming soon)
• Triangular MT: results, train, dev and test datasets
• Efficiency: all dockers and outputs, input text, and results summary
• Terminology: (coming soon)
• Unsup and very low res: (coming soon)
• Lifelong Learning: (coming soon)
• Quality Estimation: results, gold labels, submissions and baseline predictions.
• Metrics: metrics task inputs, submissions, code to compute correlations (new data for challenge sets now available!)
• Automatic Post-Editing: results, submissions, Chinese gold labels (post-edits), German gold labels (post-edits)

You can also download the past raw data from WMT07, WMT08, WMT09, WMT10, WMT11, WMT12, WMT13, WMT14, WMT15, WMT16, WMT17, WMT18, WMT19 and WMT20 .