HKUST NLP

HKUST NLP is the natural language processing research group at the Hong Kong University of Science and Technology. The group builds open-source tools, benchmarks, and datasets to advance language model research, with a focus on agent capabilities, evaluation, and reasoning. Toolathlon is one of their flagship benchmark projects, accepted at ICLR 2026. The group actively publishes code, trajectories, and leaderboards to support reproducible research in the community.