Kimi

Moonshot AI Kimi model tutorials currently published by vLLM Ascend.

Models

Model	vLLM Ascend latest guide
Kimi-K2-Thinking	Official guide
Kimi-K2.5	Official guide
Kimi-K2.6	Official guide

Serving engine

vLLMvLLM AscendSGLang

Status: Placeholder

This tab is reserved for Kimi deployment recipes on vLLM with UCM.

What to add

• Supported Kimi models and tested vLLM/UCM versions
• Docker Image or wheel installation
• UCM configuration and vllm serve command
• Verification request, expected result, and known limitations

Status: Placeholder

This tab is reserved for Kimi deployment recipes on vLLM Ascend with UCM.

What to add

• Supported Kimi models and A2/A3/A5 validation status
• Tested vLLM Ascend, CANN, and UCM versions
• UCM configuration and serving command
• Verification request, expected result, and hardware notes

Status: Placeholder

This tab is reserved for Kimi deployment recipes on SGLang with UCM.

What to add

• Supported Kimi models and tested SGLang/UCM versions
• Installation and launch command
• UCM configuration
• Verification request, expected result, and known limitations