vLLM's Kimi K3 Backend Hits 842 Tokens per Second on NVIDIA Blackwell GPUs

A unified CuTe DSL decode path on GB300 pushes Kimi K3 to 842 tokens/s at batch 1, up to 2x faster than stock vLLM.

Source

vLLM

Published

Oct 18, 2026, 1:55 AM

Author

AlphaSignal Newsroom

Read

1 min read

A unified CuTe DSL decode path on GB300 pushes Kimi K3 to 842 tokens/s at batch 1, up to 2x faster than stock vLLM.

Reporting is indexed from AlphaSignal. Rights remain with the original publisher and cited sources.

Next reads