HKGAI V3: Hong Kong's New AI Model Boosts Efficiency

Hong Kong has unveiled the latest version of its homegrown large language model, HKGAI V3. This update is a major step in building AI tools tailored to the city's industry and language. The new model improves operating efficiency and agent performance. Token compression efficiency is up more than tenfold, and uninterrupted agent runtime has increased nearly a hundredfold. Its Agent Workshop can run continuously for up to 28 hours. Developers say the model reflects Hong Kong's "cultural DNA," interpreting inputs and expressing ideas that better match local users' habits. This means more practical, locally relevant applications are expected. The model will also be open-sourced under the name ClawNet, allowing easy access for businesses and researchers. This development shows how AI is becoming more personalized and integrated into daily life.