[NVIDIA] Add Kimi K2.5 NVFP4 GB200 disaggregated vLLM benchmarks via Dynamo#1008

functionstackx merged 2 commits intomainSemiAnalysisAI/InferenceX:mainfrom

nv/kimi2.5-disagg-gb200-8k1k-1k1kSemiAnalysisAI/InferenceX:nv/kimi2.5-disagg-gb200-8k1k-1k1kCopy head branch name to clipboard