Build a world-class machine translation program
Our MT Evaluation experts will help you select the best-fit option among pre-trained models and custom models trained on your content.
Choose the right engine to save up to 70% on post-editing efforts
- Clear-cut engine analysis Receive detailed reporting on the potential of popular engines to proceed with the top performers.
- Intelligent sampling for human review Review only meaningful segments using MT evaluation metrics. Decrease evaluation data by 200 times while considering crucial model features.
- Actionable recommendations Analyze the strengths and weaknesses of different models. Train and improve the best-fit options for your language and domain.
- Fast and affordableObtain a complete evaluation and MT program strategy in less than 6 weeks.
Receive end-to-end machine translation evaluations. Relax, while we handle it all:
Data cleaning
Prepare the best data to train your machine translation engines. Your custom models will be protected from translation inconsistencies.
Model training
Apply your translation memories and glossaries to get as close as possible to human translation quality.
Test sample translations
Get test translations for your languages and domain to determine the most suitable model (from 30 different stock and custom models).
Test sample translations
Get test translations for your languages and domain to determine the most suitable model (from 30 different stock and custom models).
Model training analysis
Recognize and prevent model issues with segment-level scores including hLepor, BERTScore, and COMET.
Final analysis
Rank models based on translation quality, price, and response time. We’ll also consider translation consistency, tone-of-voice, glossaries, and more.
Model training analysis
Recognize and prevent model issues with segment-level scores including hLepor, BERTScore, and COMET.
LQA (sample review)
Compile valuable insights from your language experts through sample review. Intento’s trusted LSP partners can do this job for you as well.
Final analysis
Rank models based on translation quality, price, and response time. We’ll also consider translation consistency, tone-of-voice, glossaries, and more.
Use MT Evaluation with expert supervision for the most strategic MT program
Evaluated models
Language pairs available for evaluation