Model components / Inference code

xLLM

xLLM is an efficient LLM inference framework specifically optimized for Chinese AI accelerators. It separates a service layer that handles scheduling and availability from an engine layer that handles computation, and is deployed at scale across JD.com's core retail business.

The `xllm` package on PyPI is an unrelated project, so no package artifact is declared. Verified 2026-08-31 via GitHub, the LICENSE body and the repository README.

Openness

5 medium confidence

5.0

license: Apache-2.0(OSI)
source: public(the published repository is the engine)
core-gated: ungated(no enterprise path in the repository root and no paid build of the engine in the README)

The LICENSE body is the stock Apache-2.0 text with no appended condition, read in full rather than taken from the API's label. The repository is public and unarchived and builds the engine itself. Its root tree carries no enterprise, ee or commercial directory and its README describes no licence-gated build, so the core reads as ungated. Confidence is medium because that is a repository-and-README read rather than a pricing-page read.

https://raw.githubusercontent.com/xLLM-AI/xllm/main/LICENSE recorded 2026-08-31
LICENSE body, read in full: the stock Apache-2.0 text with no appended condition.
https://api.github.com/repos/xLLM-AI/xllm recorded 2026-08-31
Repo metadata for xLLM-AI/xllm - archived false, 1,543 stars.
https://raw.githubusercontent.com/xLLM-AI/xllm/main/README.md recorded 2026-08-31
README, read for a paid tier, an enterprise edition or a licence-gated build beside the published source, and for whether any component is stated to be private.

Adoption

2 low confidence

2.0

1,543 GitHub stars, which lands in the 1K-10K band of the stars scale. The `xllm` package on PyPI is a different project - bobazooba/xllm, an LLM finetuning library - so no package artifact is declared here and stars are the only instrument available.

https://api.github.com/repos/xLLM-AI/xllm recorded 2026-08-31
Repo metadata - stargazers_count = 1,543 - for xLLM-AI/xllm.

Capability

4 medium confidence

4.0

One band below the vllm anchor. It carries the serving shape vllm does but over a narrower hardware target - Chinese AI accelerators rather than the multi-vendor set vllm reaches - which is the same reasoning that places lmdeploy one below.

https://raw.githubusercontent.com/xLLM-AI/xllm/main/README.md recorded 2026-08-31
README, read for the engine's feature set: Service-engine decoupled serving on Chinese accelerators, with the scheduling and availability layer separated from the computation layer.

Verified 2026-08-31