Kimi K3 Set for Q3 Launch, Aims for 2.5 Trillion Parameters in Computing Race

The competition among domestic AI large models is heating up. Following the market buzz around DeepSeek V4, Moonshot AI's next-generation model, Kimi K3, has reported significant progress. According to available information, Kimi K3 is slated for an official launch in the third quarter of this year, with its parameter scale potentially reaching a staggering 2.5 trillion.
In AI, parameter count is often considered a key indicator of a model's capability. For context, the recently released DeepSeek V4 Pro has 1.6 trillion parameters, while Baidu's Wenxin 5.0 boasts approximately 2.4 trillion. This means Kimi K3 not only doubles the data volume of its predecessor, K2.X, but is also poised to surpass most mainstream domestic models and challenge the performance tier of leading global AI models.
Beyond a leap in computational power, context processing is a core strength of the Kimi series. Reports indicate the standard context length for Kimi K3 will be increased to around 1M (approximately one million words), far exceeding the 256K supported by the current K2.6 version. While internal test data suggest even greater lengths, the final context available to general users—considering the immense computational cost and operational expenses—remains to be officially confirmed.
The domestic model market is currently evolving on two parallel tracks: "cost-effectiveness" and "peak performance." On one side, DeepSeek pushes the boundaries of computational efficiency and accessibility. On the other, models like Kimi continuously advance in long-context and ultra-large-scale capabilities. The arrival of Kimi K3 will undoubtedly raise the competitive bar for domestic large models, offering users a more profound level of logical reasoning and information processing.
Related article
AI Search Mandatory Policy Fuels Exodus, DuckDuckGo Sees User Surge
Following Google's 2026 I/O conference announcement of a full AI overhaul of its search engine, many users started looking for more controllable alternatives because there was no simple "one-click disable" for AI features. The privacy-focused search
Xiaohongshu Restructures: Conan Named President, Creates AI Primary Department Dots and Overseas Division Rednote
On April 30, Xiaohongshu sent an internal memo to all employees announcing the launch of a new organizational restructuring. The core of this change involves fully integrating three business lines—community, e-commerce, and commercialization—along wi
Tencent's Xiaolongxia Surges Beyond Expectations, Team Expands Capacity 10x, Apologizes and Compensates
Tencent has officially launched WorkBuddy, an all-scenario AI intelligent agent, marking a new phase in the large model application layer race with high integration and a low deployment threshold.The product drew immediate industry attention on its l
Related Special Topic Recommendations
Comments (0)
0/500

The competition among domestic AI large models is heating up. Following the market buzz around DeepSeek V4, Moonshot AI's next-generation model, Kimi K3, has reported significant progress. According to available information, Kimi K3 is slated for an official launch in the third quarter of this year, with its parameter scale potentially reaching a staggering 2.5 trillion.
In AI, parameter count is often considered a key indicator of a model's capability. For context, the recently released DeepSeek V4 Pro has 1.6 trillion parameters, while Baidu's Wenxin 5.0 boasts approximately 2.4 trillion. This means Kimi K3 not only doubles the data volume of its predecessor, K2.X, but is also poised to surpass most mainstream domestic models and challenge the performance tier of leading global AI models.
Beyond a leap in computational power, context processing is a core strength of the Kimi series. Reports indicate the standard context length for Kimi K3 will be increased to around 1M (approximately one million words), far exceeding the 256K supported by the current K2.6 version. While internal test data suggest even greater lengths, the final context available to general users—considering the immense computational cost and operational expenses—remains to be officially confirmed.
The domestic model market is currently evolving on two parallel tracks: "cost-effectiveness" and "peak performance." On one side, DeepSeek pushes the boundaries of computational efficiency and accessibility. On the other, models like Kimi continuously advance in long-context and ultra-large-scale capabilities. The arrival of Kimi K3 will undoubtedly raise the competitive bar for domestic large models, offering users a more profound level of logical reasoning and information processing.
AI Search Mandatory Policy Fuels Exodus, DuckDuckGo Sees User Surge
Following Google's 2026 I/O conference announcement of a full AI overhaul of its search engine, many users started looking for more controllable alternatives because there was no simple "one-click disable" for AI features. The privacy-focused search
Xiaohongshu Restructures: Conan Named President, Creates AI Primary Department Dots and Overseas Division Rednote
On April 30, Xiaohongshu sent an internal memo to all employees announcing the launch of a new organizational restructuring. The core of this change involves fully integrating three business lines—community, e-commerce, and commercialization—along wi
Tencent's Xiaolongxia Surges Beyond Expectations, Team Expands Capacity 10x, Apologizes and Compensates
Tencent has officially launched WorkBuddy, an all-scenario AI intelligent agent, marking a new phase in the large model application layer race with high integration and a low deployment threshold.The product drew immediate industry attention on its l





Home






