DeepL, a German AI startup, has unveiled a new large language model (LLM) focused on translation and editing, surpassing competitors like Google Translate, ChatGPT-4, and Microsoft in translation quality. The model is powered by proprietary LLM technology, using DeepL's own training data accumulated over seven years and fine-tuned by thousands of language experts. Blind tests revealed DeepL's translations are preferred by language experts up to 2.3 times more than major competitors. DeepL's new LLM aims to set industry standards by providing more human-like translations and reducing the risk of errors, with plans to expand language support and empower businesses globally.