The Kimi K3 model API calling service has been launched on the national supercomputing internet.

PANews July 21 reports that the National Supercomputing Interconnect has launched the Kimi K3 model API call service, offering pay-as-you-go interfaces for enterprises and developers. Kimi K3 is an open-source large model with about 2.8 trillion parameters. It adopts the Kimi Delta Attention (KDA), Attention Residuals (AttnRes), and Stable LatentMoE architectures, supports a context window of up to 1 million tokens and multimodal capabilities, and ranked first in the Frontend Code Arena at the 1679th percentile. The National Supercomputing Interconnect has already integrated domestic open-source large model APIs such as Kimi-K3, DeepSeek-V4, Qwen, MiniMax-M3, and GLM-5.2, supporting capabilities including text, images, video, and speech. Users can connect with one click through the official website, without having to build their own compute clusters or deploy in complex environments.
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pinned