Thrilled to invite you to SGLang's in-person meeting with Hyperbolic on January 15th at 6 PM in 2 Embarcadero Center, San Francisco. 😻 6:00 PM - 6:30 PM: Networking with drinks and food. 6:30 PM - 6:50 PM: Ali Abdalla, HuggingFace. 6:50 PM - 7:10 PM: Jasper Zhang, CEO +

This future of AI is collaborative! Thanks so much for the great chance and night with Hyperbolic. If you want to learn and contribute more about SGLang, do not hesitate to raise your questions in SGLang's Slack or on our GitHub repo.

This future of AI is collaborative! Thanks so much for the great chance with Hyperbolic and this is also the first time I am at an SGLang in-person meet-up. 👏 👏 👏 I really love SF and the Bay Area, full of tech atmosphere and everyone is enthusiastic about LLM. 🥹🥹🥹

my personal blog on ML SYS learning journey is started by Lilian Weng, former member and leader at OpenAI security team 😂 😂 😂 github.com/zhaochenyang20… this is the link. Sorry most of the contents in Chinese, but will be more international this year! 🥲

Thanks so much for the help and for seeing the SGLang community's rapid growth.

❗️Open source MOE kernels alert❗️ Introducing COMET, a computation/communication library for MoE models from Bytedance. Battle-tested in our 10k+ GPU clusters, COMET shows promising efficiency gains and significant GPU-hour savings (millions 💰💰💰). Integration of DualPipe &

Boost your gen-AI workflow's quality by 2.8x with just $5 in 24 minutes! Check how Cognify autotunes gen-AI workflow’s quality and execution efficiency with a tiny budget in our latest blog post tinyurl.com/4tyvvdks. Paper tinyurl.com/3kx2xjn9. Code tinyurl.com/2tp9bndr.

Is a single accuracy number all we can get from model evals?🤔 🚨Does NOT tell where the model fails 🚨Does NOT tell how to improve it Introducing EvalTree🌳 🔍identifying LM weaknesses in natural language 🚀weaknesses serve as actionable guidance (paper&demo 🔗in🧵) [1/n]

After 2 months of intense work (20+ developers team, 12h shifts), we’ve fully integrated veRL-SGLang end-to-end frameworks! Thanks SGLang-RL team!

Hello, everyone! Get ready for an electrifying session this Saturday at 6 PM Pacific Time (9 PM Eastern Time) or Sunday at 9 AM Beijing Time. Our SGLang RL Team will take the stage at the SGLang Biweekly Meeting to share thrilling updates on our RLHF (Reinforcement Learning from

vLLM DeepSeek Hi, vLLM team, please stop posting these misleading benchmark results. Both sglang and trtllm can achieve much better numbers than the ones in your post. See the reproduced runs of sglang: x.com/lmsysorg/statu… This isn’t the first time you’ve shared misleading information,