Kimi releases K2.7 Code high-speed version, with output speed reaching 260 tokens/s
Summary
Kimi has launched the K2.7 Code high-speed version and opened it to Kimi Code Beta members, API developers, and Kimi Business users. The new version keeps the same model as K2.7 Code but runs 5 to 6 times faster, reaching 180 tokens per second in typical coding scenarios and up to 260 tokens per second in short-context cases. Kimi also says the model reduces average token usage by 30% while improving long-context programming instruction following and task efficiency. The pricing is set at double the original version, with higher input and output rates for standard usage.
Classifications
industries
No industries detected
applications
No applications detected
AskAI Classifications
Labels
No AI classifications detected