[Moonshot AI to Release 2.8 Trillion Parameter Kimi K3 Weights, China's Open-Source Weight Model Token Share Rises to 68%] Odaily Planet Daily reports that Chinese AI startup Moonshot AI will release the model weights for its high-performance model Kimi K3. Developers will be able to download the model, modify it for specific use cases, and run it in their own data centers or cloud environments. Kimi K3 features 2.8 trillion parameters and a 1 million token context window, enabling it to process large-scale documents and codebases in a single operation. Moonshot AI plans to subsequently publish a technical report detailing the model architecture, training methods, and performance evaluation results. Following the release of Kimi K3, Moonshot AI's daily revenue reportedly increased by at least sixfold. The company is advancing plans for a new round of funding at a $50 billion valuation and is considering going public in Hong Kong as early as this year. Bloomberg Intelligence stated that after the release of Kimi K3 and Z.AI's GLM-5.2, the share of China's open-source weight models in overall token usage has risen to 68%. AWS Bedrock, Microsoft Azure Foundry, and Google Vertex AI have yet to provide support for Chinese open-source weight models such as Kimi K3 and GLM-5.2.