Kimi K3 launches on SGLang with 423 tok/s and 11 cloud partners
ying11231 · x · 2026-07-28
- Kimi K3 is now live on SGLang with a reported 423 tok/s day-0 speed on GSM8K.
- The post highlights that the demo video was generated by Kimi K3 itself, running on the same serving stack being announced.
- Engineering details include a 2.8T MoE model, a novel KDA architecture, and deep serving optimizations such as fused decode kernels, DP attention, DSpark, PD disaggregation, and KDA-aware prefix caching.
- The ecosystem launch included 11 cloud partners serving the model from day one, framed as a sign of a mature open-source stack where model, infra, and cloud providers ship together.
Related event: SGLang Day-0 Support for Kimi K3 Boosts Throughput to 423 tok/s(8 posts)→
More from Infra
- Random Engram Reads Kill SSDs: NVLink-Backed CPU Memory and Redis Eyed for Sparse Memory Lookup — bookwormengr · 2026-09-11
- Engram's random reads don't suit SSDs; CPU-memory over NVLink could serve all 72 GPUs — bookwormengr · 2026-09-11
- 80% of the DIY LLM inference hype posters have already quit — it's brutally hard systems work — abhijithneil · 2026-09-11
- Hugging Face's Ultra Scale Playbook: a free book on training LLMs on GPU clusters — mdancho84 · 2026-09-11
- Is inference latency becoming the biggest bottleneck for production AI agents? — Euphoric_Sea632 · 2026-09-11
- LLM Serving Metrics Thread: Why TPOT and Uptime Make or Break User Experience — abhijithneil · 2026-09-11