![]() | This article has multiple issues. Please help improve it or discuss these issues on the talk page . (Learn how and when to remove these messages)
|
Kimi | |
---|---|
![]() | |
Developer(s) | Moonshot AI |
Initial release | October 2023 |
Stable release | Kimi-K2-0905-Instruct / September 5, 2025 |
Platform | |
Type | Chatbot Large language model |
License | Proprietary MIT License (Kimi-VL, Kimi-Dev) Modified MIT License (Kimi K2) |
Website | kimi kimi |
Kimi is an artificial intelligence (AI) chatbot and series of large language models developed by Moonshot AI. Its first version, released in 2023, was known for its ability to handle 200,000 Chinese characters in a single prompt. [1] In July 2025, Kimi K2 was released which performed well on several popular benchmarks and also became the top-rated open source AI model on LMArena. [2] [3] [4] In September 2025, an updated version of K2, Kimi-K2-0905-Instruct, was released. [5] [6]
Moonshot AI was founded in March 2023. In October 2023, the company officially released the Kimi chatbot and began closed-beta testing. [7]
On 16 November 2023, Kimi was released to the general public in China based on the Moonshot model. The first version of Kimi supported lossless context input of roughly 200,000 Chinese characters [1] , making it the first AI model that was capable of accepting inputs of this size. [8]
In March 2024, Moonshot AI began closed beta testing of an updated version of Kimi with a 2 million character context window. [9]
In July 2024, Kimi's “context caching” feature entered public beta. [10]
On 11 October 2024, Kimi Explore Edition, featuring AI-powered autonomous search, went live in China [11] ; monthly active users have since exceeded 36 million. [12] [13]
In November 2024, Kimi began internal testing of an AI video generation model. [14]
On 20 January 2025, Kimi K1.5 was released. Moonshot AI claimed it matched the performance of OpenAI o1 in mathematics, coding, and multimodal reasoning capabilities. [15]
In April 2025, Kimi-VL, an open source 16 billion parameter mixture of experts (MoE) large language model with 3 billion active parameters, was released. In June, a version of the model capable of reasoning was also released. [16] [17]
In June 2025, Kimi-Dev, a 72B parameter coding-focued model based on Qwen2.5-72B, was released. It achieved state of the art performance among open source models on the SWE-bench Verified benchmark. [18] [19] Also in June, Moonshot AI released Kimi-Researcher, an autonomous AI research agent available through Kimi's website and app. [20] [21]
In July 2025, Moonshot AI released Kimi K2, a 1 trillion parameter mixture of experts large language model with 32 billion active parameters which was open sourced under a modified MIT license. It achieved state of the art performance in coding benchmarks while still offering good performance in other benchmarks. [3] [4] On 9 September 2025, Moonshot AI released an updated version of K2, Kimi-K2-Instruct-0905, which improved its performance in coding tasks and increased its context window from 128K tokens to 256K tokens. [5] [6]
The Kimi chatbot and apps are free to use outside of China. However, there is a $19 USD per month subscription plan for the Kimi Researcher feature (known as "STARTER"), which gives access to 20 uses per month, high-priority access, the ability to have 2 research tasks run at the same time, and opportunities to test advanced models. In China, Kimi has six tiers of plans ranging from 5.2 yuan for four days to 399 yuan for a year of priority use. [22]