vLLM achieves 7.53s weight transfer for 1T param Kimi K2

AccBalanced · x · 2026-08-26

The vLLM team implemented a native sharded weight transfer engine using Ray Direct Transport (RDT) and NIXL. Supporting dense, MoE, and quantized models, it optimizes performance by overlapping processing stages. It achieved BF16 weight transfer for the 1T parameter Kimi K2 model in 7.53 seconds across 48 8xH100 nodes.

Related event: vLLM Ships Native Sharded Weight Transfer, Syncing Trillion-Parameter Models in Seconds(2 posts)→

Original post →

More from Infra

Infra channel →