sglang
https://github.com/sgl-project/sglang
Python
SGLang is a high-performance serving framework for large language models and multimodal models.
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported1 Subscribers
Add a CodeTriage badge to sglang
Help out
- Issues
- [Feat] [NVIDIA] DSA: Add FlashInfer GVR V2 top-k backend
- [Feat] [AMD] kimi-k3.1: PD over MoRI between same-size DCP roles
- [Feat] [Ascend] Add A3 MegaMoE support for SwiGLU/SiTU and BF16/INT experts
- [Fix] parser: split reasoning on last think end token
- [Fix] dsv4: Keep trtllm-gen sparse MLA from reading lengths the launch right before it wrote
- [Feat] model_config: Allow external models to opt into MLA caching
- [Bug] Checkpoint FP8 KV-cache scales are ignored: attention built without quant_config (GLM-4.5, Llama 4, Qwen3-MoE, GPT-OSS, HunYuan, HY3, K-EXAONE, LFM2)
- [Bug] Reasoning parser splits on first occurrence of closing think tag: quoted </think> inside reasoning truncates reasoning_content and leaks the rest of the CoT into content
- [Bug] Model gateway silently drops `cache_salt` on chat completions
- [Fix] quantization: Pad the Marlin N tile for compressed-tensors WNA16
- Docs
- Python not yet supported