SGLang Update: Adds Support for Kimi K3 and Local Video Generation
ying11231 · x · 2026-08-11
SGLang has released version v0.5.17 with several major updates:
- New Model Support: Integrated Kimi K3 with optimized prefix caching and speculative decoding, supporting 1M context with recipes for both NVIDIA and AMD GPUs. Added MiniMax-H3, an open video-and-audio generation model capable of local generation on 2x RTX 5090.
- Performance: The frontend server got an initial Rust rewrite; MoE prefill is up to 1.92x faster.
- Local AI Tracking: Added a 'Local AI' section to track progress across platforms like SM120, DGX Spark, Halo, and Apple MLX.
Related event: SGLang v0.5.17 Adds Support for Kimi K3 and Video Generation(2 posts)→
More from Infra
- Microsoft Plans to 'Significantly' Increase Production of Next-Gen AI Chips — thoefler · 2026-08-11
- fal Signs 3 Hot GenAI Model Companies, Expands H200 and B300 Capacity — gorkem · 2026-08-11
- Alphabet Aims to Raise $25B in Bonds to Fund AI Infrastructure Buildout — Beth_Kindig · 2026-08-11
- Nvidia reportedly partners with financial firms to finance $500B in AI chips and data centers — Polymarket · 2026-08-11
- Fixing Intel Arc VRAM Issues: Developer Releases Stability Patch for ComfyUI — Valuable-Subject-274 · 2026-08-11
- Oz-FP4: Emulating FP64 DGEMM on Low-Precision FP4 Tensor Cores — teortaxesTex · 2026-08-11