Launched in April 2006 as a statistical machine translation service, it used United Nations and European Parliament documents and transcripts to collect linguistic data. Instead of translating languages directly, it first translates the text into English and then pivots to the target language for most language combinations it puts in its grid, with a few exceptions including Catalan-Spanish. During a translation, it looks for patterns in millions of documents to help decide which words to choose and how to organize them in the target language. Its accuracy, which has been criticized and ridiculed on several occasions, has been measured to vary greatly between languages. In November 2016, Google announced that Google Translate would switch to a neural machine translation engine – Google Neural Machine Translation (GNMT) – which translates "entire sentences at once, rather than just part by part. It uses this context broader to help it discover the most relevant translation, which it then rearranges and adjusts to be more like a human speaking with proper grammar." Originally only enabled for a few languages in 2016, GNMT is now used in all 133 languages on the Google Translate list as of May 2022.