Unlike Google Advanced v3 API, Google AutoML does not support batch translation. Therefore, when the CAT/TMS system sends an array of segments to translate, in case of Google AutoML they will be translated one-by-one, resulting in up to 10x slower translation in Trados and memoQ.
This also means that certain Google limits will be depleted faster. In addition, Google Advanced v3 API enables you adding your custom terminology on a top of both custom and stock models, which lots of our clients find very useful. We suggest you switch to Google Advanced API for all your requests to ensure faster and more reliable translation.