Manila Times
Upholding Truth. Empowering the Philippines
Monday, Mar 23, 2026
The Nation's Story is Told with Honor
Moonshot AI Unveils Kimi K2: A New Open-Source AI Model
The Chinese startup Moonshot AI launches its Kimi K2 model, showcasing advancements in open-source AI technology.
Moonshot AI, a Beijing-based startup, has announced the launch of its new open-source artificial intelligence model, Kimi K2.
The model has been designed to excel in areas including frontier knowledge, mathematics, coding, and general agentic tasks.
Moonshot aims to maintain a competitive edge against rival companies in the rapidly evolving AI landscape, including DeepSeek and others.
Kimi K2 employs a mixture-of-experts (MoE) architecture, boasting a total of 1 trillion parameters, with 32 billion activated parameters dedicated to specific computational tasks.
This architecture allows the model to divide its functions among distinct sub-networks or 'experts,' optimizing efficiency during both pre-training and inference processes, according to information shared by the company.
The startup has released two versions of Kimi K2 in an effort to cater to different user needs.
The first version, Kimi-K2-Base, is optimized for researchers and developers seeking extensive control over fine-tuning and customization of the model.
The second version, Kimi-K2-Instruct, has undergone post-training to enhance its usability in general-purpose chat and agentic AI functions.
Kimi K2 is now available through both web and mobile applications, reinforcing Moonshot’s commitment to open-source development.
This latest release is part of a broader trend in the artificial intelligence sector, where open-source models have gained traction as a means of increasing efficiency and driving wider adoption across both startups and established tech firms.
The open-source format allows developers to access the program's source code, enabling them to modify and share its designs, repair issues, or enhance performance capabilities.
The trend towards open-source AI is gathering momentum, impacting not only companies like Moonshot AI but also larger tech corporations such as Baidu and Alibaba Cloud, which are seeking to leverage open-source methodologies to expedite innovation and accessibility in AI technologies.
The model has been designed to excel in areas including frontier knowledge, mathematics, coding, and general agentic tasks.
Moonshot aims to maintain a competitive edge against rival companies in the rapidly evolving AI landscape, including DeepSeek and others.
Kimi K2 employs a mixture-of-experts (MoE) architecture, boasting a total of 1 trillion parameters, with 32 billion activated parameters dedicated to specific computational tasks.
This architecture allows the model to divide its functions among distinct sub-networks or 'experts,' optimizing efficiency during both pre-training and inference processes, according to information shared by the company.
The startup has released two versions of Kimi K2 in an effort to cater to different user needs.
The first version, Kimi-K2-Base, is optimized for researchers and developers seeking extensive control over fine-tuning and customization of the model.
The second version, Kimi-K2-Instruct, has undergone post-training to enhance its usability in general-purpose chat and agentic AI functions.
Kimi K2 is now available through both web and mobile applications, reinforcing Moonshot’s commitment to open-source development.
This latest release is part of a broader trend in the artificial intelligence sector, where open-source models have gained traction as a means of increasing efficiency and driving wider adoption across both startups and established tech firms.
The open-source format allows developers to access the program's source code, enabling them to modify and share its designs, repair issues, or enhance performance capabilities.
The trend towards open-source AI is gathering momentum, impacting not only companies like Moonshot AI but also larger tech corporations such as Baidu and Alibaba Cloud, which are seeking to leverage open-source methodologies to expedite innovation and accessibility in AI technologies.
AI Disclaimer: An advanced artificial intelligence (AI) system generated the content of this page on its own. This innovative technology conducts extensive research from a variety of reliable sources, performs rigorous fact-checking and verification, cleans up and balances biased or manipulated content, and presents a minimal factual summary that is just enough yet essential for you to function as an informed and educated citizen. Please keep in mind, however, that this system is an evolving technology, and as a result, the article may contain accidental inaccuracies or errors. We urge you to help us improve our site by reporting any inaccuracies you find using the "Contact Us" link at the bottom of this page. Your helpful feedback helps us improve our system and deliver more precise content. When you find an article of interest here, please look for the full and extensive coverage of this topic in traditional news sources, as they are written by professional journalists that we try to support, not replace. We appreciate your understanding and assistance.
