vllm-project/vllm-torchtpu

| Documentation | Blog | User Forum | Developer Slack (#sig-tpu) |

vLLM TorchTPU is the PyTorch-native TPU backend for vLLM. It is an out-of-tree vLLM platform plugin, packaged as vllm-torchtpu, that runs vLLM's standard PyTorch model definitions on Google Cloud TPUs through TorchTPU (torch_tpu, via torch.compile(backend="tpu")) and a library of custom JAX/Pallas kernels for attention, Mixture-of-Experts, and quantized matmul.

The goal is simple: the same vLLM you already use — same CLI, same OpenAI-compatible server, same model definitions — running at TPU speed with no model code changes.

• PyTorch first. Models run through vLLM's native PyTorch runner; no re-implementation in another framework is required to serve on TPU.
• TPU-tuned kernels. Ragged Paged Attention, dense and sparse MLA, Gated DeltaNet, fused MoE (including SparseCore offload), and FP8/INT8/MXFP4 quantized matmuls written in Pallas.
• Scale-out built in. TP, DP, PP, EP, and prefill/decode context parallelism compose across single-host and multi-host TPU slices, with disaggregated prefill/decode serving.
• Standard vLLM experience. Speculative decoding, prefix caching, multi-LoRA, structured output, and multimodal inputs work through the normal vLLM flags.

See Cloud TPU Setup for project setup, quota, and provisioning commands.

vllm-torchtpu requires Python 3.12 on a Cloud TPU VM. Install vllm-torchtpu from PyPI and a matching vLLM with a single command:

pip install \
  --extra-index-url https://download.pytorch.org/whl/cpu \
  vllm-torchtpu

Or pull the pre-built container from Docker Hub (vllm/vllm-torchtpu):

docker pull vllm/vllm-torchtpu:latest

For uv, editable source installs, cross-repo development against a local vllm checkout, and Docker images, see the full Installation guide and Docker guide.

Start an OpenAI-compatible server exactly as you would on any other vLLM platform:

vllm serve meta-llama/Llama-3.2-1B-Instruct \
  --tensor-parallel-size 1 \
  --max-model-len 4096

Then send a request:

curl http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "meta-llama/Llama-3.2-1B-Instruct",
    "messages": [{"role": "user", "content": "Explain how TPUs accelerate matrix multiplication."}],
    "max_tokens": 128
  }'

The Quickstart covers offline batch inference with vllm.LLM, multi-device serving with TP/EP, and multimodal examples.

vllm-torchtpu can serve most PyTorch models supported by vLLM, with first-class TPU kernel, quantization, and parallelism support for Qwen3 / Qwen3.5, DeepSeek-V3 / R1 / V4, Kimi-K2.5 / K3, GLM-5, Llama 3 / 4, GPT-OSS, Gemma 3 / 4, Mixtral, Qwen-VL and Gemma VL families, and diffusion models via vllm-omni.

• Supported Models & Features — validated model families, feature status, and quantization formats.
• Nightly support matrices — TPU v7x unit, correctness, and benchmark status, updated automatically from nightly CI.

• TPU v7x (Ironwood) serving recipes
• TPU v6e (Trillium) serving recipes
• Kubernetes deployment
• Multi-host serving · Parallelism & scaling · Disaggregated serving (P/D)
• Quantization · Speculative decoding

We welcome contributions of all kinds — bug reports, model and feature requests, documentation fixes, and code.

• 🐞 Report bugs or request features: open a new issue.
• 👀 Review pull requests: help us keep quality high by reviewing open PRs.
• 💻 Write code: start with the Contributing Guide, which covers environment setup, DCO sign-off, the ready-label CI gate, and the PR lifecycle. Good first issues are a great place to begin.

Developer references: Testing & CI · Profiling · Benchmarking · Coding guidance · Optimization best practices

• For technical questions and feature requests, open a GitHub issue.
• For discussion with fellow users, use the TPU support topic on the vLLM Forum.
• For coordinating contributions and development, join the vLLM Developer Slack (#sig-tpu).