Samsung Sets New Benchmark with TrueBench for AI Productivity in Multilingual Workspaces

Samsung Electronics Co. has launched a new platform designed to evaluate the productivity of artificial intelligence (AI) models in workplace settings, enhancing the current industry standards. Dubbed TrueBench, this initiative aims to furnish companies with a set of comprehensive metrics to monitor the productivity of large language models (LLMs) in realistic work environments.

The Korean technology giant emphasized the increasing demand for reliable productivity measurement tools as enterprises progressively adopt AI for task completion. Existing benchmarks predominantly focus on overall performance and are mostly centered around English, restricted to single-round Q&A structures. TrueBench addresses these limitations by evaluating work-related tasks, such as content generation and data analysis, across 10 categories and 46 subcategories.

TrueBench supports 12 languages, including Korean, English, and Japanese, broadening its accessibility and applicability across various global enterprises. This diverse linguistic support is a significant advancement for companies operating in multilingual environments, facilitating more accurate and relevant productivity assessments of AI models.

Cheun Kyung-whoon, CTO of Samsung's Device eXperience Division, highlighted the company's commitment to leveraging deep industry expertise and competitive advantages gained from extensive AI research. He expressed optimism that TrueBench would establish new standards for productivity evaluation, solidifying Samsung's leadership position in the technology sector.