SGLang
High-performance serving framework for LLM and VLM workloads.
Stories
Filter storiesPoolside released Laguna S 2.1, a 118B-parameter open-weight MoE with 1M context and SGLang day-zero support. Poolside and partners cite SWE-bench, Terminal-Bench, and local-agent tests.
Tencent released Hy3 with 21B active parameters, a 256K context window, BF16/FP8 weights, and day-one vLLM/SGLang support. Kilo Code, Nous Portal, and OpenRouter also made it free for limited windows.
LMSYS and Modal shipped DFlash plus Spec V2 in SGLang, claiming 4.3x baseline throughput and 1.5x native MTP on Qwen3.5-397B-A17B. It cuts latency and serving cost for very large open models.
SGLang v0.5.12 added native DeepSeek V4 support with ShadowRadix prefix caching, HiSparse CPU-extended KV, MegaMoE kernels, and Blackwell MLA work. The release broadens hardware targets and improves long-context serving efficiency for open runtimes.
OpenBMB released MiniCPM-V 4.6 1.3B, claiming 55.8 percent lower vision-encoding FLOPs, 75.7 ms TTFT on a 4090, and about 1.5x token throughput over Qwen3.5 0.8B. It targets edge deployment across mobile platforms and common inference stacks.
OpenClaw beta added live control of a real Chrome session through Chrome DevTools MCP; the project also added native SGLang provider support and parallel tool calling work. Try it if you need self-hosted agents to handle authenticated browser flows with local inference backends.