AMD MI355X matches Nvidia B300 on Kimi K3 serving, with 24.4% more throughput

AIvLLM says eight AMD MI355X GPUs deliver 24.4% more throughput than an HGX B300 on Kimi K3 serving, at 7,567 versus 6,083 tokens per second per GPU at 50 tokens per second per user. The post also reports 87.4% more tokens per dollar on SemiAnalysis's AgentX agentic coding benchmark.

vLLMOfficial

@vllm_project

View on X

🚀 @AMD MI355X just showed competitive performance to NVIDIA B300 on @Kimi_Moonshot K3 serving!

8× MI355X delivers 24.4% more throughput than HGX B300 (7,567 vs 6,083 tok/s/GPU) at 50 tok/s/user on @SemiAnalysis_ AgentX agentic coding benchmark, and 87.4% more tokens per dollar.

We cover the vLLM optimizations, AMD kernels, and hardware details that closed the gap, plus patterns for spotting bottlenecks so you can get more out of vLLM on your own GPUs.

Thanks @AMD @embeddedLLM @SemiAnalysis_

Full writeup: 🔗 https://vllm.ai/blog/2026-10-16-kimi-k3-mi355x

Source details

vLLM · x.comPublished Oct 16, 5:34 PM ET · added here 5:50 PM ET

Original title: 🚀 @AMD MI355X just showed competitive performance to NVIDIA B300 on @Kimi_Moonshot K3 serving!

AMD MI355X matches Nvidia B300 on Kimi K3 serving, with 24.4% more throughput

Topics