Select Language:
On September 5th, Fenghuang Web Tech announced the release of a new version of its Kimi K2 model, aimed at enhancing performance in real-world programming tasks.
The updated Kimi K2 boasts significant improvements, including a jump in its context length from 128K to an impressive 256K. This expansion allows the model to better handle complex, lengthy projects, providing developers with more robust support for long-term coding tasks. Additionally, the model has upgraded its Agentic Coding capabilities, demonstrating superior performance in both public benchmarking tests and actual programming scenarios.
A noteworthy feature of this release is the introduction of a high-speed API version, which supports output speeds of up to 60 to 100 tokens per second, catering to users demanding faster interactions.
Fenghuang Web Tech confirmed that the Kimi platform now hosts the new Kimi-K2-0905-Preview model API, featuring the expanded 256K context length. Pricing remains consistent with the previous 0711 version, offering existing users a seamless upgrade path without additional costs.
Images accompanying the announcement highlight the model’s capabilities and improvements, emphasizing its enhanced speed and extended context support.
ChatGPT
Perplexity
Gemini AI
Grok AI
