Recently, Moonshot AI released the latest update about the Kimi K2 Turbo version of the AI model. After the efforts of engineers, the output speed of the Kimi K2 turbo-preview model has significantly improved, with a stable output speed now reaching 60 Tokens per second, and a maximum of up to 100 Tokens per second. This advancement marks a significant improvement in the model's efficiency when processing data, allowing users to experience smoother service when using it.
Kimi K2 is a foundational model launched by Moonshot AI, based on the MoE (Mixture of Experts) architecture, featuring strong code generation capabilities and excellent performance in general agent tasks. The total parameter count of the model reaches 1 trillion, with 32 billion active parameters. Kimi K2 was officially released on August 1st, and initially ran at a speed of 10 Tokens per second. After optimization, it has been increased to 40 Tokens per second, and further enhanced to reach a stable output of 60 Tokens per second.
To attract more users, the Kimi K2 Turbo version currently offers a special discount of 50% off. The input cost of the model is divided into two scenarios: when the cache hits, the input price is 2 yuan per million Tokens, and when the cache does not hit, it is 8 yuan, while the output price is 32 yuan. However, it should be noted that this promotional offer will end on September 1st, after which the price will return to its original level.
Kimi K2 not only excels in speed but also demonstrates formidable technical strength. Its power enables it to perform well in various tasks, especially in code generation and complex agent tasks, where it has received widespread recognition and positive feedback from users. Moonshot AI stated that they will continue to optimize the model's performance, striving to provide users with more efficient and accurate AI services.
Key Points:
✨ The output speed of the Kimi K2 Turbo version AI model has been upgraded to 60 Tokens per second, with a maximum of up to 100 Tokens per second.
💰 The model is currently enjoying a 50% discount, with reasonable input and output pricing to attract user experience.
🚀 Kimi K2 performs exceptionally well in code generation and general agent tasks, receiving positive reviews from users, and will continue to be optimized in the future.
