This repository contains a group of BentoML example projects, showing you how to serve and deploy open-source Large Language Models using vLLM, a high-throughput and memory-efficient inference engine. Every model directory contains the code to add OpenAI compatible endpoints to the BentoML Service.

💡 You can use these examples as bases for advanced code customization, such as custom model, inference logic or vLLM options. For simple LLM hosting with OpenAI compatible endpoints without writing any code, see OpenLLM.

See here for a full list of BentoML example projects.

The following is an example of serving one of the LLMs in this repository: Llama 3.1 8B Instruct.

Prerequisites

• If you want to test the Service locally, we recommend you use an Nvidia GPU with at least 16G VRAM.
• Gain access to the model in Hugging Face.

Install dependencies

```bash git clone https://github.com/bentoml/BentoVLLM.git cd BentoVLLM/llama3.1-8b-instruct

Recommend UV and Python 3.11

uv venv && uv pip install -r requirements.txt

export HF_TOKEN= ```

Run the BentoML Service

We have defined a BentoML Service in service.py. Run bentoml serve in your project directory to start the Service.

```bash $ bentoml serve .

2024-01-18T07:51:30+0800 [INFO] [cli] Starting production HTTP BentoServer from "service:VLLM" listening on http://localhost:3000 (Press CTRL+C to quit) INFO 01-18 07:51:40 modelrunner.py:501] Capturing the model for CUDA graphs. This may lead to unexpected consequences if the model is not static. To run the model in eager mode, set 'enforceeager=True' or use '--enforce-eager' in the CLI. INFO 01-18 07:51:40 modelrunner.py:505] CUDA graphs can take additional 1~3 GiB memory per GPU. If you are running out of memory, consider decreasing `gpumemoryutilization` or enforcing eager mode. INFO 01-18 07:51:46 modelrunner.py:547] Graph capturing finished in 6 secs. ```

The server is now active at http://localhost:3000. You can interact with it using the Swagger UI or in other different ways.

OpenAI-compatible endpoints

```python from openai import OpenAI

client = OpenAI(baseurl='http://localhost:3000/v1', apikey='na')

completion = client.chat.completions.create( model=client.models.list().data[0].id, messages=[ { "role": "user", "content": "Who are you? Please respond in pirate speak!" } ], stream=True, ) for chunk in completion: # Extract and print the content of the model's reply print(chunk.choices[0].delta.content or "", end="") ```

These OpenAI-compatible endpoints also support vLLM extra parameters. For example, you can force the chat completion output a JSON object by using the guided_json parameters:

```python from openai import OpenAI

client = OpenAI(baseurl='http://localhost:3000/v1', apikey='na')

json_schema = { "type": "object", "properties": { "city": {"type": "string"} } }

completion = client.chat.completions.create( model=client.models.list().data[0].id, messages=[ { "role": "user", "content": "What is the capital of France?" } ], extrabody=dict(guidedjson=json_schema), ) print(completion.choices[0].message.content) # will return something like: {"city": "Paris"} ```

All supported extra parameters are listed in vLLM documentation.

Note: If your Service is deployed with protected endpoints on BentoCloud, you need to set the environment variable OPENAI_API_KEY to your BentoCloud API key first.

bash export OPENAI_API_KEY={YOUR_BENTOCLOUD_API_TOKEN}

You can then use the following line to replace the client in the above code snippet. Refer to Obtain the endpoint URL to retrieve the endpoint URL.

python client = OpenAI(base_url='your_bentocloud_deployment_endpoint_url/v1')

For detailed explanations of the Service code, see vLLM inference.

Deploy to BentoCloud

After the Service is ready, you can deploy the application to BentoCloud for better management and scalability. Sign up if you haven't got a BentoCloud account.

Make sure you have logged in to BentoCloud.

bash bentoml cloud login

Create a BentoCloud secret to store the required environment variable and reference it for deployment.

```bash bentoml secret create huggingface HFTOKEN=$HFTOKEN

bentoml deploy . --secret huggingface ```

Note: For custom deployment in your own infrastructure, use BentoML to generate an OCI-compliant image.

Featured models

In addition to Llama 3.1 8B Instruct, we also have examples for other models in the subdirectories of this repository:

| Model | Links | |-------|-------| | deepseek-prover-v2-671b | GitHub • Hugging Face | | deepseek-r1-671b | GitHub • Hugging Face | | deepseek-v3-671b | GitHub • Hugging Face | | devstral-small-2505 | GitHub • Hugging Face | | gemma3-4b-instruct | GitHub • Hugging Face | | gpt-oss-120b | GitHub • Hugging Face | | gpt-oss-20b | GitHub • Hugging Face | | jamba1.6-large | GitHub • Hugging Face | | jamba1.6-mini | GitHub • Hugging Face | | llama3.1-8b-instruct | GitHub • Hugging Face | | llama3.1-8b-kv-offloading | GitHub • Hugging Face | | llama3.2-3b-instruct | GitHub • Hugging Face | | llama3.3-70b-instruct | GitHub • Hugging Face | | llama4-17b-maverick-instruct | GitHub • Hugging Face | | llama4-17b-scout-instruct | GitHub • Hugging Face | | magistral-small-2506 | GitHub • Hugging Face | | mistral-small-3.1-24b-instruct-2503 | GitHub • Hugging Face | | phi4-14b | GitHub • Hugging Face | | phi4-14b-reasoning | GitHub • Hugging Face | | phi4-14b-reasoning-plus | GitHub • Hugging Face | | qwen3-235b-a22b | GitHub • Hugging Face | | qwen3-30b-a3b | GitHub • Hugging Face | | qwen3-8b | GitHub • Hugging Face | | qwen3-coder-480b-a35b | GitHub • Hugging Face | | r1-0528-qwen3-8b | GitHub • Hugging Face | | r1-distill-llama3.3-70b | GitHub • Hugging Face |