- DeepL, a vanguard in Language AI, has launched an avant-garde language model, meticulously crafted for translation and editing tasks.
- It is equipped with a tailored Large Language Model, seven years of proprietary data, and expert linguistic tutoring, resulting in more fluent and accurate translations.
OUR TAKE
“Bringing our new LLM to market is a significant milestone for our research team, however, the real proof lies in the tangible impact we can bring to our customers When testing our new model, professional translators noted that DeepL translations need far fewer manual edits to achieve better quality than competitors. For businesses who translate a large volume of documents, this truly impacts their efficiency and productivity, meaning team members can focus on tasks with high-business value,” says Stefan Mesken, VP Research of DeepL.
–Vicky Wu, BTW reporter
What happened
DeepL, a vanguard in Language AI, has launched an avant-garde language model, meticulously crafted for translation and editing tasks. This leap forward, grounded in a specialised Large Language Model (LLM) technology, represents a considerable stride in the progression of DeepL’s Language AI platform for commercial entities, recalibrating industry standards for translation quality and efficiency.
The innovative model consolidates three unique attributes: a precisely calibrated LLM for language, proprietary data amassed over seven years for content creation and translation purposes, and human-led model tutoring by accomplished linguistic professionals. These elements in unison guarantee translations that echo human fluency, curtailing errors and misinformation.
The firm also executed blind tests between the new and the classic LLM it uses. They demonstrated a 1.7x accuracy uplift with the new one for combinations between English, Japanese, and Simplified Chinese, alongside a 1.4x improvement in the English-German pair. Compared with prevalent translation utilities like Google Translate and ChatGPT-4, DeepL’s model was preferred 1.3 and 1.7 times more respectively, assuring users diminished time spent on post-translation editing.
Also read: Microsoft’s AI system ‘SpreadsheetLLM’ elevates firm productivity
Also read: BNP Paribas partners with Mistral AI to implement LLMs
Why it’s important
The release comes on the heels of a substantial growth spurt for DeepL. The startup has expanded its portfolio of languages, inaugurated its maiden U.S. office, and debuted novel business-oriented products, including the DeepL Write Pro AI assistant. In May, DeepL secured a new investment of $300 million, reaching a $2 billion valuation. The company has been prompt in leveraging the acquired funds, asserting that its latest model represents a substantial upgrade over its preceding translation systems. The recent outcomes of this innovative solution notably elevate the standards for AI translation quality. Blind tests have indicated that DeepL’s translations are selected by language experts at a rate up to 2.3 times higher compared to those produced by leading AI competitors.
“This is only the beginning of DeepL’s superior LLM-powered Language AI solutions for businesses,” said Jarek Kutylowski, CEO and Founder, DeepL. “We are a research-driven company to our core and are investing heavily in pushing the boundaries of translation and writing quality, efficiency, and adaptability. Our goal is to empower our customers worldwide with the industry-leading technology they need to thrive and scale globally without a language barrier standing in the way.”
