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
- fix(model-gateway): accept token ID prompts in /v1/completions
- [Bug] Radix cache `evictable_size()` double-counts shared pages, killing the scheduler at idle
- fix: boot Inkling on Hopper without an explicit attention backend
- [Spec] Support dp attention for DFLASH speculative decoding
- [Bug] Qwen3.6/Qwen3.8 multimodal mRoPE positions are passed to a 1D fused QK RMSNorm+RoPE CUDA kernel
- [Bug] llama3 tool-call parser silently deletes any leading JSON object from message content
- [Bug] 8 tool-call parsers give wrong or missing calls in token-by-token streaming
- [mem_cache][11/N] refactor: extract KVCache and BaseSWAKVPool into pool/base.py
- fix: pass anthropic effort xhigh through instead of collapsing to max
- [Bug] Anthropic endpoint: tool_reference parts in tool_result break chat templates without deferred-reference support (HTTP 500)
- Docs
- Python not yet supported