Moonshot Launches Kimi K3, the World's Largest Open-Weight Model at 2.8 Trillion Parameters

My take: Moonshot AI just launched Kimi K3, the largest open-weight model in existence: 2.8 trillion parameters, Mixture-of-Experts architecture, a 1-million-token context window, and native multimodal capability. The API is already available, and the full weights drop on July 27.

The detail that matters most is not the model size: it is that the weights are open. That means any technical team can audit it in depth, deploy it on their own infrastructure, or fine-tune it with their own data without depending on a third party's terms of service. For organizations with compliance or privacy requirements, the difference between a closed model and an open-weight one can be decisive.

Worth noting: all performance figures published so far come from Moonshot itself, the team that built the model. That does not mean the numbers are wrong, but it is worth waiting for independent evaluations before treating them as a definitive benchmark.

Does your team have a clear policy on when to use closed models versus open-weight models?