Get in the driver's seat of machine translation in 2021

  • The Machine Translation market is accelerating. Since last year, 13 more vendors have offered pre-trained stock models. 
  • In this study, we have evaluated 29 MT engines – 14 more than a year ago.
  • The language coverage is unprecedented and approaches 100,000 language pairs – more than a 6x increase since last year.
  • We have identified 19 MT engines as leaders across 13 language pairs and 7 industry sectors. Learn which ones!
  • For the first time, we have evaluated several open-source pre-trained MT models.
  • This report is based on automated similarity scoring. It will be augmented by human LQA for 3 language pairs (coming soon)

Discover which MT engines are best suited for your language pair

The report evaluates general-purpose Cloud Machine Translation systems with pre-trained translation models available via API

commercial machine translation engines
 
 Commercial MT Engines
  • Alibaba (eCommerce model)
  • Alibaba General
  • Amazon Translate
  • Apptek
  • Baidu Translate
  • DeepL
  • Elia
  • Globalese
  • Google Advanced Translation
  • GTCom YeeCloud
  • IBM Watson
  • Microsoft Text Translator
commercial machine translation engines
 
 Commercial MT Engines 2
  • ModernMT Realtime
  • Naver Papago
  • Kawamura / NICT
  • Pangeanic
  • PROMT
  • Rozetta T-400
  • SYSTRAN PNMT
  • Tilde
  • Tencent Cloud TMT
  • Yandex Translate
  • Youdao
  • XL8
open source machine translation engines

 

Open-Source MT Engines

  • M2M-100
  • mBART50
  • OPUS MT

Analyze MT performance across 7 industries

  • Education
  • Finance
  • Healthcare
  • Hospitality
  • Legal
  • Entertainment
  • General
machine translation performance across different industries
machine translation evaluation scores

Find out which scores to consider in the MT evaluation

We compare the latest semantic similarity scores with a high correlation to human judgment (COMET, BERTScore, PRISM) and conventional metrics (hLEPOR).

Get reliable and data-driven insights from an industry leader

Group 402-2

Gartner names Intento a "Cool Vendor 2021" in Conversational and Natural Language Technologies.

 

Group 402-2

We cooperate with all players in the language industry – MT providers, LSPs, and TMS/CAT vendors.

 

Group 402-2

Intento has trained and evaluated 691 models across 51 language pairs for global companies in 2021.

 

e2f's partnership with Intento provides the best-fit machine translation solutions for our forward thinking clients. Using Intento's intelligent MT selection process, e2f is able to deliver much higher quality translations in much less time than when using an off the shelf MT solution.

 

image-2-1
Robert Lo Bue
VP Localization | e2f

Download the report to get your exclusive 2021 MT cheatsheet
The State of Machine Translation
MAIN Logo White BG