Google Cloud adds Day 0 SGLang support for Moonshot’s 2.8T Kimi K3 on GKE
ying11231 · x · 2026-07-29
Google Cloud has announced Day 0 support for Moonshot’s Kimi K3 with SGLang.
The guide shows how to serve the 2.8T-parameter MoE model on GKE, including DSPARK speculative decoding. It emphasizes that models at this scale are primarily an infrastructure problem: Google’s A4 and A4X VMs provide the memory bandwidth and interconnect needed to keep K3 fast under real concurrency.
Google outlines three deployment paths:
- Model Garden
- AI Hypercomputer recipes
- GKE
The post includes a full deployment guide in the comments.
More from Infra
- NVIDIA launches RTX Spark for local AI on Windows PCs — ostrisai · 2026-10-16
- SkyRL ships Day-0 RL support for Beam with +62% LoRA decode throughput in vLLM — vllm_project · 2026-10-16
- DFlash2 delivers up to 3.03x decode-only speedup for Beam across five benchmarks — vllm_project · 2026-10-16
- vLLM's Beam serving kernels cut GB200 forward time from 8.15 to 6.47 ms at TP8 batch 128 — vllm_project · 2026-10-16
- Reflection's 501B MoE Beam hits 463 tok/s on 8x GB200 with NVFP4 at day 0 — zhyncs42 · 2026-10-16
- Gemma 4 31B QAT on one 24GB card: 46 tok/s plain, ~80 with draft model on RTX 4090 — Subject_Poet_3861 · 2026-10-16