Warning

A.X fork — build this tree's CUDA kernels; a Python-only install silently drops fixes.

This branch changes compiled code, not only Python. Notably csrc/libtorch_stable/cache_kernels.cu fixes the fp8_ds_mla KV-cache writer on SM100 (B200 / B300): FlashMLA's sparse fp8 decode kernel reads each tile scale back as E8M0 rounded toward zero, so the writer stores the scale as a power of two. Without it, --kv-cache-dtype fp8 on FLASHMLA_SPARSE has ~67% attention-output error instead of ~6%, which shows up as wrong digits, phone numbers and tool-call arguments.

Install path	Kernel fixes included?
pip install -e . (full source build)	✅ yes
VLLM_USE_PRECOMPILED=1 pip install -e .	❌ no — fetches upstream binaries
pip install vllm==0.23.0 + copy this branch's vllm/*.py over it	❌ no — the .so files are stock
build_ext rebuild that swaps only _C / _moe_C	❌ no — the writer lives in _C_stable_libtorch

When overlaying Python onto the stock wheel (common in Dockerfiles), rebuild the extensions from this tree and copy all of them:

TORCH_CUDA_ARCH_LIST="10.0a" python setup.py build_ext --inplace   # 10.3a for B300
for f in _C _C_stable_libtorch _moe_C cumem_allocator; do cp -f vllm/$f.abi3.so "$SITE_PKG/$f.abi3.so"; done

Then check the installed binary, not the source tree (0 means stock kernels):

strings "$(python -c 'import os, vllm; print(os.path.dirname(vllm.__file__))')/_C_stable_libtorch.abi3.so" | grep -c VLLM_DS_MLA_UE8M0_SCALE

Only fp8_ds_mla (--kv-cache-dtype fp8 with FLASHMLA_SPARSE) on compute capability 10.x is affected; the default bf16 KV cache is not. VLLM_DS_MLA_UE8M0_SCALE=0/1 forces the fix off/on (default: on for SM100). It is read by the compiled kernel, so it does nothing on stock binaries.

Easy, fast, and cheap LLM serving for everyone

| Documentation | Blog | Paper | Twitter/X | User Forum | Developer Slack |

🔥 We have built a vLLM website to help you get started with vLLM. Please visit vllm.ai to learn more. For events, please visit vllm.ai/events to join us.

---

About

vLLM is a fast and easy-to-use library for LLM inference and serving.

Originally developed in the Sky Computing Lab at UC Berkeley, vLLM has grown into one of the most active open-source AI projects built and maintained by a diverse community of many dozens of academic institutions and companies from over 2000 contributors.

vLLM is fast with:

• State-of-the-art serving throughput
• Efficient management of attention key and value memory with PagedAttention
• Continuous batching of incoming requests, chunked prefill, prefix caching
• Fast and flexible model execution with piecewise and full CUDA/HIP graphs
• Quantization: FP8, MXFP8/MXFP4, NVFP4, INT8, INT4, GPTQ/AWQ, GGUF, compressed-tensors, ModelOpt, TorchAO, and more
• Optimized attention kernels including FlashAttention, FlashInfer, TRTLLM-GEN, FlashMLA, and Triton
• Optimized GEMM/MoE kernels for various precisions using CUTLASS, TRTLLM-GEN, CuTeDSL
• Speculative decoding including n-gram, suffix, EAGLE, DFlash
• Automatic kernel generation and graph-level transformations using torch.compile
• Disaggregated prefill, decode, and encode

vLLM is flexible and easy to use with:

• Seamless integration with popular Hugging Face models
• High-throughput serving with various decoding algorithms, including parallel sampling, beam search, and more
• Tensor, pipeline, data, expert, and context parallelism for distributed inference
• Streaming outputs
• Generation of structured outputs using xgrammar or guidance
• Tool calling and reasoning parsers
• OpenAI-compatible API server, plus Anthropic Messages API and gRPC support
• Efficient multi-LoRA support for dense and MoE layers
• Support for NVIDIA GPUs, AMD GPUs, and x86/ARM/PowerPC CPUs. Additionally, diverse hardware plugins such as Google TPUs, Intel Gaudi, IBM Spyre, Huawei Ascend, Rebellions NPU, Apple Silicon, MetaX GPU, and more.

vLLM seamlessly supports 200+ model architectures on Hugging Face, including:

• Decoder-only LLMs (e.g., Llama, Qwen, Gemma)
• Mixture-of-Expert LLMs (e.g., Mixtral, DeepSeek-V3, Qwen-MoE, GPT-OSS)
• Hybrid attention and state-space models (e.g., Mamba, Qwen3.5)
• Multi-modal models (e.g., LLaVA, Qwen-VL, Pixtral)
• Embedding and retrieval models (e.g., E5-Mistral, GTE, ColBERT)
• Reward and classification models (e.g., Qwen-Math)

Find the full list of supported models here.

Getting Started

Install vLLM with uv (recommended) or pip:

uv pip install vllm

Or build from source for development.

Visit our documentation to learn more.

• Installation
• Quickstart
• List of Supported Models

Contributing

We welcome and value any contributions and collaborations. Please check out Contributing to vLLM for how to get involved.

Citation

If you use vLLM for your research, please cite our paper:

@inproceedings{kwon2023efficient,
  title={Efficient Memory Management for Large Language Model Serving with PagedAttention},
  author={Woosuk Kwon and Zhuohan Li and Siyuan Zhuang and Ying Sheng and Lianmin Zheng and Cody Hao Yu and Joseph E. Gonzalez and Hao Zhang and Ion Stoica},
  booktitle={Proceedings of the ACM SIGOPS 29th Symposium on Operating Systems Principles},
  year={2023}
}

Contact Us

• For technical questions and feature requests, please use GitHub Issues
• For discussing with fellow users, please use the vLLM Forum
• For coordinating contributions and development, please use Slack
• For security disclosures, please use GitHub's Security Advisories feature
• For collaborations and partnerships, please contact us at [email protected]

Media Kit

• If you wish to use vLLM's logo, please refer to our media kit repo