Kimi Releases K2.7 Code HighSpeed Mode with Up to 6x Faster Inference
English summary
Moonshot AI has introduced a high-speed mode for its open-source multimodal coding model Kimi K2.7 Code. The new mode achieves up to 6× faster inference, delivering around 180 tokens per second on coding tasks with median-length inputs and up to 260 tokens per second on shorter-context tasks. The HighSpeed mode is currently rolling out to participants in the Kimi Code Beta Program, Kimi API developers, and Kimi Business users, though access remains limited due to capacity constraints. No invitation is needed; anyone joining the Beta Program can gain access. The company states it will continue improving the model and expanding access as capacity grows.
Chinese summary
月之暗面（Moonshot AI）为其开源多模态编程模型 Kimi K2.7 Code 推出了高速模式。新模式实现了最高 6 倍的推理速度提升，在中等长度输入的编程任务上可达到约 180 tokens/秒，在短上下文任务上最高可达 260 tokens/秒。该高速模式正逐步向 Kimi Code Beta 计划成员、Kimi API 开发者和 Kimi 商业用户开放，但因容量限制，访问目前仍有限。无需邀请，加入 Beta 计划就有机会获得访问权限。公司表示随着容量增加，将继续优化模型并扩大访问范围。
Key points
The HighSpeed mode enables up to 6× faster inference, with speeds of ~180 tok/s on median-length coding tasks and up to 260 tok/s on shorter-context tasks.
高速模式实现最高 6 倍推理加速，中等长度编程任务约 180 tokens/秒，短上下文任务最高 260 tokens/秒。
Access is initially limited to Kimi Code Beta Program members, API developers, and Kimi Business users due to capacity constraints.
由于容量限制，访问目前仅对 Kimi Code Beta 计划成员、API 开发者和商业用户开放。
No invitation is required; anyone can join the Beta Program to gain access, and the company plans to expand access as capacity allows.
无需邀请，加入 Beta 计划即可获得访问权限；公司计划随着容量增加逐步扩大访问范围。