HKGAI V3: Hong Kong's AI Model Gets Major Upgrade

Hong Kong has unveiled the latest version of its homegrown large language model, HKGAI V3. This model aims to create AI tools specifically for the city's industries and language. What's interesting is the significant improvements in performance. Token compression efficiency has increased more than tenfold, and uninterrupted agent runtime is nearly a hundred times better than the previous version. The HKGAI V3's Agent Workshop can run continuously for up to 28 hours in a single session. Developers say the model embeds Hong Kong’s "cultural DNA" to better reflect local users' habits. This means it can interpret language and express ideas in a more culturally attuned way. The Secretary for Innovation, Technology and Industry, Sun Dong, stated that Hong Kong is moving towards the forefront of AI innovation with this release. He expects the model to enable more practical, locally relevant applications, including enhancements to HKChat and HKPilot, the city’s AI chatbot and document processing tool. The model will also be open-sourced under the name ClawNet. This aims to lower barriers for businesses and research institutions to develop customized AI agents. This development means Hong Kong is actively shaping AI to fit its unique needs and culture.