Source: https://seerepo.com/compare/NVIDIA/TensorRT-LLM/vs/vllm-project/vllm
Markdown source: https://seerepo.com/compare/NVIDIA/TensorRT-LLM/vs/vllm-project/vllm.md
Description: TensorRT-LLM and vllm share 28 technologies: FastAPI, pyzmq, OpenAI, starlette, NVIDIA, quack-kernels, sentencepiece, tiktoken, torch, triton, transformers, einops, safetensors, torchvision, datasets, MCP, peft, pillow, jinja2, Pydantic, msgspec, jsonschema, opentelemetry-api, psutil, prometheus-client, OpenTelemetry, opentelemetry-sdk, and prometheus-fastapi-instrumentator. Both are written in Python.

# TensorRT-LLM vs vllm

TensorRT-LLM and vllm share 28 technologies: FastAPI, pyzmq, OpenAI, starlette, NVIDIA, quack-kernels, sentencepiece, tiktoken, torch, triton, transformers, einops, safetensors, torchvision, datasets, MCP, peft, pillow, jinja2, Pydantic, msgspec, jsonschema, opentelemetry-api, psutil, prometheus-client, OpenTelemetry, opentelemetry-sdk, and prometheus-fastapi-instrumentator. Both are written in Python.

- [NVIDIA/TensorRT-LLM](https://seerepo.com/r/NVIDIA/TensorRT-LLM): TensorRT LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and supports state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT LLM also contains components to create Python and C++ runtimes that orchestrate the inference execution in a performant way.
- [vllm-project/vllm](https://seerepo.com/r/vllm-project/vllm): A high-throughput and memory-efficient inference and serving engine for LLMs

## What they share and what sets them apart

- In common: FastAPI, pyzmq, OpenAI, starlette, NVIDIA, quack-kernels, sentencepiece, tiktoken, torch, triton, transformers, einops, safetensors, torchvision, datasets, MCP, peft, pillow, jinja2, Pydantic, msgspec, jsonschema, opentelemetry-api, psutil, prometheus-client, OpenTelemetry, opentelemetry-sdk, prometheus-fastapi-instrumentator
- Only in TensorRT-LLM: blobfile, h5py, anyio, uvicorn, Rich, matplotlib, Plotly, accelerate, apache-tvm-ffi, cuda-python, flash-linear-attention, flashinfer-python, nvidia-cudnn-frontend, openai-codex, openai-harmony, optimum, rouge, rouge-score, torch-c-dlpack-ext, torchao, diffusers, flash-attn, xgrammar, librosa, mistral-common, partial-json-parser, tritonclient, evaluate, llguidance, lm-eval, jsonlines, mako, opentelemetry-semantic-conventions-ai, boto3
- Only in vllm: axum, gitdb, Hugging Face, Tokio, Ray, AIOHTTP, tower, mkdocs-material, indicatif, torchaudio, tokenizers, markdown, markupsafe, pymdown-extensions, image, pydantic-core, python-json-logger, tracing, tracing-subscriber

## Side by side

| | TensorRT-LLM | vllm |
|---|---|---|
| Size | 2,884,778 lines written, 3,120 code files | 2,375,117 lines written, 3,289 code files |
| Who writes it | 1.7% of its commits signed by AI. With Claude Code, CodeRabbit, GitHub Copilot, and Cursor | 12.8% of its commits signed by AI. With Claude Code, Codex, Gemini CLI, Cursor, GitHub Copilot, Devin, Mistral Vibe, CodeRabbit, Factory, and Qwen Code |
| Activity | Last push November 8, 2026, 10,997 commits, 15,038 stars | Last push November 4, 2026, 24,593 commits, 96,302 stars |
| Tests | 2,703 test files. 58.6% of its code files have a test that imports them | 2,419 test files. 60.2% of its code files have a test that imports them |

- Size: Lines written by people and agents, without generated ones; code files, without tests.
- Who writes it: A commit counts as AI's when it carries its signature: a Co-Authored-By line, its author name or the merge of its branch.
- Activity: From GitHub and its history, on the day SeeRepo read it.
- Tests: The files that test the code, and the share of code files a test imports.

## Frequently asked questions

### Which is bigger, TensorRT-LLM or vllm?

TensorRT-LLM, with 2,884,778 lines written; vllm has 2,375,117.

The proof: [NVIDIA/TensorRT-LLM's code at commit b433473](https://github.com/NVIDIA/TensorRT-LLM/tree/b433473ba0d0d49764f4d2fb3d45c82716b946dc)

### Which has more GitHub stars, TensorRT-LLM or vllm?

vllm, with 96,302 stars; TensorRT-LLM has 15,038.

The proof: [vllm-project/vllm on GitHub](https://github.com/vllm-project/vllm)

### Which has more tests, TensorRT-LLM or vllm?

TensorRT-LLM, with 2,703 test files; vllm has 2,419.

The proof: [NVIDIA/TensorRT-LLM's code at commit b433473](https://github.com/NVIDIA/TensorRT-LLM/tree/b433473ba0d0d49764f4d2fb3d45c82716b946dc)

### Which was updated more recently, TensorRT-LLM or vllm?

TensorRT-LLM: its last push was on November 8, 2026, and vllm's on November 4, 2026.

The proof: [NVIDIA/TensorRT-LLM's commits on GitHub](https://github.com/NVIDIA/TensorRT-LLM/commits)

### Which has more commits signed by AI, TensorRT-LLM or vllm?

vllm: an AI signed 12.8% of its commits, with Claude Code, Codex, Gemini CLI, Cursor, GitHub Copilot, Devin, Mistral Vibe, CodeRabbit, Factory, and Qwen Code; in TensorRT-LLM, 1.7%.

The proof: [the commits each tool signed in vllm-project/vllm](https://seerepo.com/r/vllm-project/vllm/ai)

## Sources

- NVIDIA/TensorRT-LLM: Read on November 8, 2026, at commit b433473 (https://github.com/NVIDIA/TensorRT-LLM/tree/b433473ba0d0d49764f4d2fb3d45c82716b946dc)
- vllm-project/vllm: Read on November 4, 2026, at commit 75f6761 (https://github.com/vllm-project/vllm/tree/75f67614971a44050140e3422c1a65fe5004a72b)

## What is your repo built with?

Any public repo, with no account: https://seerepo.com/r/owner/repo/built, with the owner and name it has on GitHub.