PANews, July 17 – As reported by Xinhua News Agency, on the eve of the upcoming 2026 World Artificial Intelligence Conference and High-Level Conference on Global AI Governance, the next-generation model Kimi K3 was released on the 16th, with a parameter scale reaching 2.8 trillion. This is currently the world’s largest open-source model by parameter count, marking a new step forward in the development of China’s AI models. Beijing Moonshot AI, the company behind the release, introduced that Kimi K3 natively supports visual understanding, features a 1-million-token context window, and is optimized for complex task scenarios such as software engineering, knowledge work, deep research, and multimodal understanding, further enhancing the model’s processing capability for complex tasks.
It is understood that in evaluations, Kimi K3’s overall intelligence level approaches that of the world’s leading closed-source models. A relevant person in charge at Moonshot AI stated that during the training of the new-generation Kimi K3 foundation model, the company used a self-developed underlying model architecture and accumulated a complete set of scientific model training methodologies.
