Twitxy Saved

LMSYS Org

Verified account This account is verified. Learn more

@lmsysorg

Large Model Systems Organization: We developed SGLang @sgl_project (sglang.io), Chatbot Arena (now @arena), and Vicuna!
US lmsys.org Joined August 2024
204 Following 18K Followers 1.3K Posts

Latest posts 5

Pinned post

LMSYS Org Verified account @lmsysorg
πŸ—“οΈ The SGLang Summit Community Event Calendar is now live! Explore 20+ community events around the Summit and plan your week in SF.

πŸ›  Workshops: Get hands-on with new tools and technologies, learn directly from experts, and pick up practical skills you can put to work.

πŸ’» Technical Meetups & Demos: See what teams across the AI stack are building, exchange ideas, and get into the technical details with the engineers behind the work.

🀝 Happy Hours & Community Gatherings: Catch up with fellow builders, meet new people, and continue conversations outside the conference halls.

πŸ“Nov 12-13, Fort Mason
πŸ”— Register now πŸ‘‡ luma.com/sglsummit26-ev…
luma.com SGLang Summit 2026 Community Events Β· Events Calendar View and follow events from SGLang Summit 2026 Community Events on Luma. Community and partner events around SGLang Summit 2026 (Nov 12–13, SF). Hosting dinners, meetups, networking, or just plain...
4 replies 5 reposts 28 likes 2.3K views View on X
LMSYS Org Verified account @lmsysorg
Breakout sessions at SGLang Summit 2026 bring together teams working on the systems around agents, from evaluation and experiments to messaging channels and serving.

Explore the sessions 🧡

πŸ“ Nov 12–13 Β· Fort Mason, San Francisco
Breakout access included with expo pass: luma.com/sglang-summit
LMSYS Org Verified account @lmsysorg
Check out the breakout sessions at SGLang Summit 2026 covering how teams build and scale training systems for models and agents!

πŸ“ Nov 12–13 Β· Fort Mason, San Francisco
Breakout access included with expo pass: luma.com/sglang-summit

Explore select sessions below πŸ‘‡
More sessions to be announced
2 replies 0 reposts 10 likes 1.2K views View on X
LMSYS Org Verified account @lmsysorg
XM-Sparse: Faster Decode with SGLang HiSparse

@netpreme builds memory systems that give GPUs more capacity. Its X-Memβ„’ uses a high-bandwidth interconnect to fetch offloaded KV faster than host DRAM over PCIe, with more capacity for larger batches.

- Faster decode: 49.25% / 63.58% lower time per token at 20% / 40% KV misses, versus host DRAM. Same 16K context and batch 192/GPU.
- Higher throughput: 2.09–3.04Γ— host DRAM across 16K–1M contexts at 20% misses, using each setup’s best tested batch.

Read the full blog below πŸ‘‡
9 replies 7 reposts 52 likes 3.5K views View on X