Top AI Repos — open-source AI, indexed and scored

Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.

runpod-workers/worker-vllm

The Runpod worker template for serving our large language model endpoints. Powered by vLLM.

83MITPython

View on GitHub

Stars: 460
Forks: 384
Watchers: 460
Open issues: 36
Contributors: 31

Star history, last 90 days+1 over 6 snapshots

2026-07-24 · 4592026-08-06 · 460

Star history, last 90 days
Date	Stars
2026-07-24	459
2026-07-25	459
2026-07-28	460
2026-07-30	460
2026-07-31	460
2026-08-06	460

Today

— stars today

This week

— stars this week

This month

— stars this month

Momentum

15.0

growth rate 0.00%/day

From the README

OpenAI-Compatible vLLM Serverless Endpoint Worker

Deploy OpenAI-Compatible Blazing-Fast LLM Endpoints powered by the vLLM Inference Engine on Runpod Serverless with just a few clicks.

Current vLLM version: 0.26.0

Check out our Load Balancer implementation here: vLLM Load Balancer

Setting up the Serverless Worker
Usage: OpenAI Compatibility
Usage: Standard (Non-OpenAI)
- Request Input Parameters
- Sampling Parameters
  - Text Input Formats

Setting up the Serverless Worker

Option 1: Deploy Any Model Using Pre-Built Docker Image [Recommended]

🚀 Deploy Guide: Follow our step-by-step deployment guide to deploy using the Runpod Console.

📦 Docker Image: runpod/worker-v1-vllm:<version>

Available Versions: See GitHub Releases
CUDA Compatibility: Inherits the CUDA runtime of the vllm/vllm-openai base image used at build time.

Configuration

Configure worker-vllm using environment variables:

Environment Variable	Description	Default	Options
`MODEL_NAME`	Path of the model weights		Local folder or Hugging Face repo ID
`HF_TOKEN`	HuggingFace access token for gated/private models		Your HuggingFace access token
`MAX_MODEL_LEN`	Model's maximum context length		Integer (e.g., 4096)
`QUANTIZATION`	Quantization method		"awq", "gptq", "squeezellm", "bitsandbytes"
`TENSOR_PARALLEL_SIZE`	Number of GPUs	1	Integer
`GPU_MEMORY_UTILIZATION`	Fraction of GPU memory to use	0.9	Float between 0.0 and 1.0
`MAX_NUM_SEQS`	Maximum number of sequences per iteration	256	Integer
`CUSTOM_CHAT_TEMPLATE`	Custom chat template override		Jinja2 template string
`ENABLE_AUTO_TOOL_CHOICE`	Enable automatic tool selection	false	boolean (true or false)
`TOOL_CALL_PARSER`	Parser for tool calls		"mistral", "hermes", "llama3_json", "granite", "deepseek_v3", etc.
`OPENAI_SERVED_MODEL_NAME_OVERRIDE`	Override served model name in API		String
`MAX_CONCURRENCY`	Maximum concurrent requests	30	Integer

Pass any vLLM engine arg as an environment variable: the worker translates env vars into vllm serve CLI flags. Any env var whose name is the UPPERCASED form of a vllm serve flag is applied automatically:

Environment Variable	vLLM CLI Flag	Example Value
`MAX_MODEL_LEN`	`--max-model-len`	`4096`
`ENFORCE_EAGER`	`--enforce-eager`	`true`
`ENABLE_CHUNKED_PREFILL`	`--enable-chunked-prefill`	`true`
`SPECULATIVE_CONFIG`	`--speculative-config`	`'{"model": "org/draft", "num_speculative_tokens": 3}'`

Backward-compat aliases are also honored: MODEL_NAME (--model), MODEL_REVISION (--revision), TOKENIZER_NAME (--tokenizer), OPENAI_SERVED_MODEL_NAME_OVERRIDE (--served-model-name), CUSTOM_CHAT_TEMPLATE (--chat-template).

Values are passed straight through to vLLM (ints, floats, JSON blobs all work). For anything not recognized — e.g. a flag newer than the worker's list — use the ultimate escape hatch, whose contents are appended verbatim to the vllm serve command (and override earlier settings):

VLLM_EXTRA_ARGS="--override-generation-config '{\"max_new_tokens\": 512}' --uvicorn-log-level warning"

Configuration File (config.yaml)

As an alternative to environment variables, you can supply a complete vllm serve --config file using the CLI key names (hyphens or underscores both work):

model: meta-llama/Llama-3.1-8B-Instruct
max-model-len: 8192
gpu-memory-utilization: 0.90
quantization: awq
tensor-parallel-size: 2

Mount the file anywhere into the container and point the VLLM_CONFIG_FILE env var at it. CLI flags built from environment variables take precedence over config file values (standard vllm serve behavior).

For the complete list of all available environment variables, examples, and detailed descriptions: Configuration

Option 2: Build Docker Image with Model Inside

To build an image with the model baked in, you must specify the following docker arguments when building the image.

Prerequisites

Docker

Arguments

Required
- MODEL_NAME
Optional
- MODEL_REVISION: Model revision to load (default: main).
- VLLM_VERSION: Tag of the official vllm/vllm-openai base image to use (default: v0.23.0).
- BASE_PATH: Storage directory where huggingface cache and model will be located. (default: /runpod-volume, which will utilize network storage if you attach it or create a local directory within the image if you don't. If your intention is to bake the model into the image, you should set this to something like /models to make sure there are no issues if you were to accidentally attach network storage.)
- QUANTIZATION
- TOKENIZER_NAME: Tokenizer repository if you would like to use a different tokenizer than the one that comes with the model. (default: None, which uses the model's tokenizer)
- TOKENIZER_REVISION: Tokenizer revision to load (default: main).

For the remaining settings, you may apply them as environment variables when running the container. Supported environment variables are listed in the Environment Variables section.

Example: Building an image with OpenChat-3.5

docker build -t username/image:tag --build-arg MODEL_NAME="openchat/openchat_3.5" --build-arg BASE_PATH="/models" .

Example: Choosing the vLLM Version

The image is based on the official vllm/vllm-openai image; switching vLLM versions is a single build arg:

docker build -t username/image:tag --build-arg VLLM_VERSION=v0.10.0 --build-arg MODEL_NAME="meta-llama/Llama-3.1-8B-Instruct" .

To run a nightly vLLM build, point VLLM_VERSION at a nightly tag or use latest at your own risk.

(Optional) Including Huggingface Token

If the model you would like to deploy is private or gated, you will need to include it during build time as a Docker secret, which will protect it from being exposed in the image and on DockerHub.

Enable Docker BuildKit (required for secrets).

export DOCKER_BUILDKIT=1

Export your Hugging Face token as an environment variable

export HF_TOKEN="your_token_here"

Add the token as a secret when building

docker build -t username/image:tag --secret id=HF_TOKEN --build-arg MODEL_NAME="openchat/openchat_3.5" .

Compatible Model Architectures

You can deploy any model on Hugging Face that is supported by vLLM. For the complete and up-to-date list of supported model architectures, see the vLLM Supported Models documentation.

Usage: OpenAI Compatibility

The vLLM Worker is fully compatible with OpenAI's API, and you can use it with any OpenAI Codebase by changing only 3 lines in total. The supported routes are Chat Completions, Models, Responses, and Messages - with both streaming and non-streaming.

Modifying your OpenAI Codebase to use your deployed vLLM Worker

Python (similar to Node.js, etc.):

When initializing the OpenAI Client in your code, change the api_key to your Runpod API Key and the base_url to your Runpod Serverless Endpoint URL in the following format: https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1, filling in your deployed endpoint ID. For example, if your Endpoint ID is abc1234, the URL would be https://api.runpod.ai/v2/abc1234/openai/v1.
- Before:
```
from openai import OpenAI

client = OpenAI(api_key=os.environ.get("OPENAI_API_KEY"))
```
- After:
```
from openai import OpenAI

client = OpenAI(
    api_key=os.environ.get("RUNPOD_API_KEY"),
    base_url="https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1",
)
```

Change the model parameter to your deployed model's name whenever using Completions or Chat Completions.

Before:

response = client.chat.completions.create(
    model="gpt-3.5-turbo",
    messages=[{"role": "user", "content": "Why is Runpod the best platform?"}],
    temperature=0,
    max_tokens=100,
)

After:

response = client.chat.completions.create(
    model="<YOUR DEPLOYED MODEL REPO/NAME>",
    messages=[{"role": "user", "content": "Why is Runpod the best platform?"}],
    temperature=0,
    max_tokens=100,
)

Using http requests:

Change the Authorization header to your Runpod API Key and the url to your Runpod Serverless Endpoint URL in the following format: https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1

Before:

curl https://api.openai.com/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $OPENAI_API_KEY" \
-d '{
"model": "gpt-4",
"messages": [
  {
    "role": "user",
    "content": "Why is Runpod the best platform?"
  }
],
"temperature": 0,
"max_tokens": 100
}'

After:

curl https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer <YOUR OPENAI API KEY>" \
-d '{
"model": "<YOUR DEPLOYED MODEL REPO/NAME>",
"messages": [
  {
    "role": "user",
    "content": "Why is Runpod the best platform?"
  }
],
"temperature": 0,
"max_tokens": 100
}'

OpenAI Request Input Parameters:

When using the chat completion feature of the vLLM Serverless Endpoint Worker, you can customize your requests with the following parameters:

Chat Completions [RECOMMENDED]

Supported Chat Completions Inputs and Descriptions

Parameter	Type	Default Value	Description
`messages`	Union[str, List[Dict[str, str]]]		List of messages, where each message is a dictionary with a `role` and `content`. The model's chat template will be applied to the messages automatically, so the model must have one or it should be specified as `CUSTOM_CHAT_TEMPLATE` env var.
`model`	str		The model repo that you've deployed on your Runpod Serverless Endpoint. If you are unsure what the name is or are baking the model in, use the guide to get the list of available models in the Examples: Using your Runpod endpoint with OpenAI section
`temperature`	Optional[float]	0.7	Float that controls the randomness of the sampling. Lower values make the model more deterministic, while higher values make the model more random. Zero means greedy sampling.
`top_p`	Optional[float]	1.0	Float that controls the cumulative probability of the top tokens to consider. Must be in (0, 1]. Set to 1 to consider all tokens.
`n`	Optional[int]	1	Number of output sequences to return for the given prompt.
`max_tokens`	Optional[int]	None	Maximum number of tokens to generate per output sequence.
`seed`	Optional[int]	None	Random seed to use for the generation.
`stop`	Optional[Union[str, List[str]]]	list	List of strings that stop the generation when they are generated. The returned output will not contain the stop strings.
`stream`	Optional[bool]	False	Whether to stream or not
`presence_penalty`	Optional[float]	0.0	Float that penalizes new tokens based on whether they appear in the generated text so far. Values > 0 encourage the model to use new tokens, while values < 0 encourage the model to repeat tokens.
`frequency_penalty`	Optional[float]	0.0	Float that penalizes new tokens based on their frequency in the generated text so far. Values > 0 encourage the model to use new tokens, while values < 0 encourage the model to repeat tokens.
`logit_bias`	Optional[Dict[str, float]]	None	Unsupported by vLLM
`user`	Optional[str]	None	Unsupported by vLLM

Examples: Using your Runpod endpoint with OpenAI

First, initialize the OpenAI Client with your Runpod API Key and Endpoint URL:

from openai import OpenAI
import os

# Initialize the OpenAI Client with your Runpod API Key and Endpoint URL
client = OpenAI(
    api_key=os.environ.get("RUNPOD_API_KEY"),
    base_url="https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1",
)

Chat Completions:

This is the format used for GPT-4 and focused on instruction-following and chat. Examples of Open Source chat/instruct models include meta-llama/Llama-2-7b-chat-hf, mistralai/Mixtral-8x7B-Instruct-v0.1, openchat/openchat-3.5-0106, NousResearch/Nous-Hermes-2-Mistral-7B-DPO and more. However, if your model is a completion-style model with no chat/instruct fine-tune and/or does not have a chat template, you can still use this if you provide a chat template with the environment variable CUSTOM_CHAT_TEMPLATE.

Streaming:

# Create a chat completion stream
response_stream = client.chat.completions.create(
    model="<YOUR DEPLOYED MODEL REPO/NAME>",
    messages=[{"role": "user", "content": "Why is Runpod the best platform?"}],
    temperature=0,
    max_tokens=100,
    stream=True,
)
# Stream the response
for response in response_stream:
    print(chunk.choices[0].delta.content or "", end="", flush=True)

Non-Streaming:

# Create a chat completion
response = client.chat.completions.create(
    model="<YOUR DEPLOYED MODEL REPO/NAME>",
    messages=[{"role": "user", "content": "Why is Runpod the best platform?"}],
    temperature=0,
    max_tokens=100,
)
# Print the response
print(response.choices[0].message.content)

Getting a list of names for available models:

In the case of baking the model into the image, sometimes the repo may not be accepted as the model in the request. In this case, you can list the available models as shown below and use that name.

models_response = client.models.list()
list_of_models = [model.id for model in models_response]
print(list_of_models)

OpenAI Responses API

Path: /openai/v1/responses (full URL: https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1/responses)

Supports the OpenAI Responses API request shape. Like other /openai/ routes, this is served directly—use the /openai/ prefix rather than the RunPod native job queue for these calls.

{
  "model": "meta-llama/Llama-3.1-8B-Instruct",
  "input": "Tell me a joke."
}

Using HTTP requests:

curl https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <YOUR RUNPOD API KEY>" \
  -d '{
    "model": "<YOUR DEPLOYED MODEL REPO/NAME>",
    "input": "Tell me a joke."
  }'

Anthropic Messages API

Path: /openai/v1/messages (full URL: https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1/messages)

Supports the Anthropic Messages API format. Served directly, bypassing the RunPod queue.

{
  "model": "meta-llama/Llama-3.1-8B-Instruct",
  "max_tokens": 256,
  "messages": [
    {"role": "user", "content": "Hello!"}
  ]
}

Using HTTP requests:

curl https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1/messages \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <YOUR RUNPOD API KEY>" \
  -d '{
    "model": "<YOUR DEPLOYED MODEL REPO/NAME>",
    "max_tokens": 256,
    "messages": [
      {"role": "user", "content": "Hello!"}
    ]
  }'

Usage: Standard (Non-OpenAI)

Request Input Parameters

Click to expand table

You may either use a prompt or a list of messages as input. Under the hood, prompt is proxied to vLLM's /v1/completions and messages to /v1/chat/completions (with the model's chat template applied, so the model must have one or you must set the CUSTOM_CHAT_TEMPLATE env var).

Argument	Type	Default	Description
`prompt`	str		Prompt string to generate text based on. Proxied to `/v1/completions`.
`messages`	list[dict[str, str]]		List of messages, which will automatically have the model's chat template applied. Overrides `prompt`.
`sampling_params`	dict	{}	Sampling parameters forwarded in the request body (temperature, top_p, max_tokens, ...).
`stream`	bool	False	Whether to enable streaming of output. If True, raw SSE chunks are streamed as they are generated.

You can also call any vLLM route directly with the generic proxy form:

{
  "input": {
    "route": "/v1/chat/completions",
    "method": "POST",
    "body": {
      "model": "<YOUR DEPLOYED MODEL>",
      "messages": [{"role": "user", "content": "hello"}],
      "stream": true
    }
  }
}

Sampling Parameters

Below are all available sampling parameters that you can specify in the sampling_params dictionary. If you do not specify any of these parameters, the default values will be used.

Click to expand table

Argument	Type	Default	Description
`n`	int	1	Number of output sequences generated from the prompt. The top `n` sequences are returned.
`best_of`	Optional[int]	`n`	Number of output sequences generated from the prompt. The top `n` sequences are returned from these `best_of` sequences. Must be ≥ `n`. Treated as beam width in beam search. Default is `n`.
`presence_penalty`	float	0.0	Penalizes new tokens based on their presence in the generated text so far. Values > 0 encourage new tokens, values < 0 encourage repetition.
`frequency_penalty`	float	0.0	Penalizes new tokens based on their frequency in the generated text so far. Values > 0 encourage new tokens, values < 0 encourage repetition.
`repetition_penalty`	float	1.0	Penalizes new tokens based on their appearance in the prompt and generated text. Values > 1 encourage new tokens, values < 1 encourage repetition.
`temperature`	float	1.0	Controls the randomness of sampling. Lower values make it more deterministic, higher values make it more random. Zero means greedy sampling.
`top_p`	float	1.0	Controls the cumulative probability of top tokens to consider. Must be in (0, 1]. Set to 1 to consider all tokens.
`top_k`	int	-1	Controls the number of top tokens to consider. Set to -1 to consider all tokens.
`min_p`	float	0.0	Represents the minimum probability for a token to be considered, relative to the most likely token. Must be in [0, 1]. Set to 0 to disable.
`use_beam_search`	bool	False	Whether to use beam search instead of sampling.
`length_penalty`	float	1.0	Penalizes sequences based on their length. Used in beam search.
`early_stopping`	Union[bool, str]	False	Controls stopping condition in beam search. Can be `True`, `False`, or `"never"`.
`stop`	Union[None, str, List[str]]	None	List of strings that stop generation when produced. The output will not contain these strings.
`stop_token_ids`	Optional[List[int]]	None	List of token IDs that stop generation when produced. Output contains these tokens unless they are special tokens.
`ignore_eos`	bool	False	Whether to ignore the End-Of-Sequence token and continue generating tokens after its generation.
`max_tokens`	int	16	Maximum number of tokens to generate per output sequence.
`skip_special_tokens`	bool	True	Whether to skip special tokens in the output.
`spaces_between_special_tokens`	bool	True	Whether to add spaces between special tokens in the output.

Text Input Formats

You may either use a prompt or a list of messages as input.

prompt The prompt string can be any string, and the model's chat template will not be applied to it unless apply_chat_template is set to true, in which case it will be treated as a user message.
```
Example:
```json
{
  "input": {
    "prompt": "why sky is blue?",
    "sampling_params": {
      "temperature": 0.7,
      "max_tokens": 100
    }
  }
}
```
```

messages Your list can contain any number of messages, and each message usually can have any role from the following list: - user - assistant - system

However, some models may have different roles, so you should check the model's chat template to see which roles are required.

The model's chat template will be applied to the messages automatically, so the model must have one.

Example:

{
  "input": {
    "messages": [
      {
        "role": "system",
        "content": "You are a helpful AI assistant that provides clear and concise responses."
      },
      {
        "role": "user",
        "content": "Can you explain the difference between supervised and unsupervised learning?"
      },
      {
        "role": "assistant",
        "content": "Sure! Supervised learning uses labeled data, meaning each input has a corresponding correct output. The model learns by mapping inputs to known outputs. In contrast, unsupervised learning works with unlabeled data, where the model identifies patterns, structures, or clusters without predefined answers."
      }
    ],
    "sampling_params": {
      "temperature": 0.7,
      "max_tokens": 100
    }
  }
}

Rendered by GitHub from the current README

Read on GitHub

Top contributors

No contributor data yet — it is fetched during the contributor sync pass.

Quality score

Would you bet a product on this? Bounded 0–100 and slow moving.

83/10083

Maintenance: 90%; How recently the default branch was pushed to
Releases: 86%; Recency and cadence of tagged releases
Community: 83%; Contributor breadth and bus factor
Issue backlog: 35%; Open issues measured against the size of the audience
Documentation: 100%; README depth, plus a homepage and topics
Licensing: 100%; How freely the code can be adopted

What to know before adopting

Nothing flagged — maintained, documented and cleanly licensed.
Usable in commercial products with attribution.

Details

Created: 3 years ago
Last push: 8 days ago
Last release: last month
Release tag: v2.22.5
Releases in the last year: 30
Default branch: main
Size: 29k KB
Owner type: Organization
Owner location: United States of America
Discovered via: topic:vllm
First indexed: 13 days ago
Last synced: 12 days ago

runpod-workers/worker-vllm

The Runpod worker template for serving our large language model endpoints. Powered by vLLM.

83MITPython

View on GitHub

Stars: 460
Forks: 384
Watchers: 460
Open issues: 36
Contributors: 31

Star history, last 90 days+1 over 6 snapshots

2026-07-24 · 4592026-08-06 · 460

Star history, last 90 days
Date	Stars
2026-07-24	459
2026-07-25	459
2026-07-28	460
2026-07-30	460
2026-07-31	460
2026-08-06	460

Today

— stars today

This week

— stars this week

This month

— stars this month

Momentum

15.0

growth rate 0.00%/day

From the README

OpenAI-Compatible vLLM Serverless Endpoint Worker

Deploy OpenAI-Compatible Blazing-Fast LLM Endpoints powered by the vLLM Inference Engine on Runpod Serverless with just a few clicks.

Current vLLM version: 0.26.0

Check out our Load Balancer implementation here: vLLM Load Balancer

Setting up the Serverless Worker
Usage: OpenAI Compatibility
Usage: Standard (Non-OpenAI)
- Request Input Parameters
- Sampling Parameters
  - Text Input Formats

Setting up the Serverless Worker

Option 1: Deploy Any Model Using Pre-Built Docker Image [Recommended]

🚀 Deploy Guide: Follow our step-by-step deployment guide to deploy using the Runpod Console.

📦 Docker Image: runpod/worker-v1-vllm:<version>

Available Versions: See GitHub Releases
CUDA Compatibility: Inherits the CUDA runtime of the vllm/vllm-openai base image used at build time.

Configuration

Configure worker-vllm using environment variables:

Environment Variable	Description	Default	Options
`MODEL_NAME`	Path of the model weights		Local folder or Hugging Face repo ID
`HF_TOKEN`	HuggingFace access token for gated/private models		Your HuggingFace access token
`MAX_MODEL_LEN`	Model's maximum context length		Integer (e.g., 4096)
`QUANTIZATION`	Quantization method		"awq", "gptq", "squeezellm", "bitsandbytes"
`TENSOR_PARALLEL_SIZE`	Number of GPUs	1	Integer
`GPU_MEMORY_UTILIZATION`	Fraction of GPU memory to use	0.9	Float between 0.0 and 1.0
`MAX_NUM_SEQS`	Maximum number of sequences per iteration	256	Integer
`CUSTOM_CHAT_TEMPLATE`	Custom chat template override		Jinja2 template string
`ENABLE_AUTO_TOOL_CHOICE`	Enable automatic tool selection	false	boolean (true or false)
`TOOL_CALL_PARSER`	Parser for tool calls		"mistral", "hermes", "llama3_json", "granite", "deepseek_v3", etc.
`OPENAI_SERVED_MODEL_NAME_OVERRIDE`	Override served model name in API		String
`MAX_CONCURRENCY`	Maximum concurrent requests	30	Integer

Environment Variable	vLLM CLI Flag	Example Value
`MAX_MODEL_LEN`	`--max-model-len`	`4096`
`ENFORCE_EAGER`	`--enforce-eager`	`true`
`ENABLE_CHUNKED_PREFILL`	`--enable-chunked-prefill`	`true`
`SPECULATIVE_CONFIG`	`--speculative-config`	`'{"model": "org/draft", "num_speculative_tokens": 3}'`

VLLM_EXTRA_ARGS="--override-generation-config '{\"max_new_tokens\": 512}' --uvicorn-log-level warning"

Configuration File (config.yaml)

As an alternative to environment variables, you can supply a complete vllm serve --config file using the CLI key names (hyphens or underscores both work):

model: meta-llama/Llama-3.1-8B-Instruct
max-model-len: 8192
gpu-memory-utilization: 0.90
quantization: awq
tensor-parallel-size: 2

For the complete list of all available environment variables, examples, and detailed descriptions: Configuration

Option 2: Build Docker Image with Model Inside

To build an image with the model baked in, you must specify the following docker arguments when building the image.

Prerequisites

Docker

Arguments

Required
- MODEL_NAME
Optional
- MODEL_REVISION: Model revision to load (default: main).
- VLLM_VERSION: Tag of the official vllm/vllm-openai base image to use (default: v0.23.0).
- BASE_PATH: Storage directory where huggingface cache and model will be located. (default: /runpod-volume, which will utilize network storage if you attach it or create a local directory within the image if you don't. If your intention is to bake the model into the image, you should set this to something like /models to make sure there are no issues if you were to accidentally attach network storage.)
- QUANTIZATION
- TOKENIZER_NAME: Tokenizer repository if you would like to use a different tokenizer than the one that comes with the model. (default: None, which uses the model's tokenizer)
- TOKENIZER_REVISION: Tokenizer revision to load (default: main).

For the remaining settings, you may apply them as environment variables when running the container. Supported environment variables are listed in the Environment Variables section.

Example: Building an image with OpenChat-3.5

docker build -t username/image:tag --build-arg MODEL_NAME="openchat/openchat_3.5" --build-arg BASE_PATH="/models" .

Example: Choosing the vLLM Version

The image is based on the official vllm/vllm-openai image; switching vLLM versions is a single build arg:

docker build -t username/image:tag --build-arg VLLM_VERSION=v0.10.0 --build-arg MODEL_NAME="meta-llama/Llama-3.1-8B-Instruct" .

To run a nightly vLLM build, point VLLM_VERSION at a nightly tag or use latest at your own risk.

(Optional) Including Huggingface Token

If the model you would like to deploy is private or gated, you will need to include it during build time as a Docker secret, which will protect it from being exposed in the image and on DockerHub.

Enable Docker BuildKit (required for secrets).

export DOCKER_BUILDKIT=1

Export your Hugging Face token as an environment variable

export HF_TOKEN="your_token_here"

Add the token as a secret when building

docker build -t username/image:tag --secret id=HF_TOKEN --build-arg MODEL_NAME="openchat/openchat_3.5" .

Compatible Model Architectures

You can deploy any model on Hugging Face that is supported by vLLM. For the complete and up-to-date list of supported model architectures, see the vLLM Supported Models documentation.

Usage: OpenAI Compatibility

Modifying your OpenAI Codebase to use your deployed vLLM Worker

Python (similar to Node.js, etc.):

When initializing the OpenAI Client in your code, change the api_key to your Runpod API Key and the base_url to your Runpod Serverless Endpoint URL in the following format: https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1, filling in your deployed endpoint ID. For example, if your Endpoint ID is abc1234, the URL would be https://api.runpod.ai/v2/abc1234/openai/v1.
- Before:
```
from openai import OpenAI

client = OpenAI(api_key=os.environ.get("OPENAI_API_KEY"))
```
- After:
```
from openai import OpenAI

client = OpenAI(
    api_key=os.environ.get("RUNPOD_API_KEY"),
    base_url="https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1",
)
```

Change the model parameter to your deployed model's name whenever using Completions or Chat Completions.

Before:

response = client.chat.completions.create(
    model="gpt-3.5-turbo",
    messages=[{"role": "user", "content": "Why is Runpod the best platform?"}],
    temperature=0,
    max_tokens=100,
)

After:

response = client.chat.completions.create(
    model="<YOUR DEPLOYED MODEL REPO/NAME>",
    messages=[{"role": "user", "content": "Why is Runpod the best platform?"}],
    temperature=0,
    max_tokens=100,
)

Using http requests:

Change the Authorization header to your Runpod API Key and the url to your Runpod Serverless Endpoint URL in the following format: https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1

Before:

curl https://api.openai.com/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $OPENAI_API_KEY" \
-d '{
"model": "gpt-4",
"messages": [
  {
    "role": "user",
    "content": "Why is Runpod the best platform?"
  }
],
"temperature": 0,
"max_tokens": 100
}'

After:

curl https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer <YOUR OPENAI API KEY>" \
-d '{
"model": "<YOUR DEPLOYED MODEL REPO/NAME>",
"messages": [
  {
    "role": "user",
    "content": "Why is Runpod the best platform?"
  }
],
"temperature": 0,
"max_tokens": 100
}'

OpenAI Request Input Parameters:

When using the chat completion feature of the vLLM Serverless Endpoint Worker, you can customize your requests with the following parameters:

Chat Completions [RECOMMENDED]

Supported Chat Completions Inputs and Descriptions

Parameter	Type	Default Value	Description
`messages`	Union[str, List[Dict[str, str]]]		List of messages, where each message is a dictionary with a `role` and `content`. The model's chat template will be applied to the messages automatically, so the model must have one or it should be specified as `CUSTOM_CHAT_TEMPLATE` env var.
`model`	str		The model repo that you've deployed on your Runpod Serverless Endpoint. If you are unsure what the name is or are baking the model in, use the guide to get the list of available models in the Examples: Using your Runpod endpoint with OpenAI section
`temperature`	Optional[float]	0.7	Float that controls the randomness of the sampling. Lower values make the model more deterministic, while higher values make the model more random. Zero means greedy sampling.
`top_p`	Optional[float]	1.0	Float that controls the cumulative probability of the top tokens to consider. Must be in (0, 1]. Set to 1 to consider all tokens.
`n`	Optional[int]	1	Number of output sequences to return for the given prompt.
`max_tokens`	Optional[int]	None	Maximum number of tokens to generate per output sequence.
`seed`	Optional[int]	None	Random seed to use for the generation.
`stop`	Optional[Union[str, List[str]]]	list	List of strings that stop the generation when they are generated. The returned output will not contain the stop strings.
`stream`	Optional[bool]	False	Whether to stream or not
`presence_penalty`	Optional[float]	0.0	Float that penalizes new tokens based on whether they appear in the generated text so far. Values > 0 encourage the model to use new tokens, while values < 0 encourage the model to repeat tokens.
`frequency_penalty`	Optional[float]	0.0	Float that penalizes new tokens based on their frequency in the generated text so far. Values > 0 encourage the model to use new tokens, while values < 0 encourage the model to repeat tokens.
`logit_bias`	Optional[Dict[str, float]]	None	Unsupported by vLLM
`user`	Optional[str]	None	Unsupported by vLLM

Examples: Using your Runpod endpoint with OpenAI

First, initialize the OpenAI Client with your Runpod API Key and Endpoint URL:

from openai import OpenAI
import os

# Initialize the OpenAI Client with your Runpod API Key and Endpoint URL
client = OpenAI(
    api_key=os.environ.get("RUNPOD_API_KEY"),
    base_url="https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1",
)

Chat Completions:

Streaming:

# Create a chat completion stream
response_stream = client.chat.completions.create(
    model="<YOUR DEPLOYED MODEL REPO/NAME>",
    messages=[{"role": "user", "content": "Why is Runpod the best platform?"}],
    temperature=0,
    max_tokens=100,
    stream=True,
)
# Stream the response
for response in response_stream:
    print(chunk.choices[0].delta.content or "", end="", flush=True)

Non-Streaming:

# Create a chat completion
response = client.chat.completions.create(
    model="<YOUR DEPLOYED MODEL REPO/NAME>",
    messages=[{"role": "user", "content": "Why is Runpod the best platform?"}],
    temperature=0,
    max_tokens=100,
)
# Print the response
print(response.choices[0].message.content)

Getting a list of names for available models:

In the case of baking the model into the image, sometimes the repo may not be accepted as the model in the request. In this case, you can list the available models as shown below and use that name.

models_response = client.models.list()
list_of_models = [model.id for model in models_response]
print(list_of_models)

OpenAI Responses API

Path: /openai/v1/responses (full URL: https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1/responses)

Supports the OpenAI Responses API request shape. Like other /openai/ routes, this is served directly—use the /openai/ prefix rather than the RunPod native job queue for these calls.

{
  "model": "meta-llama/Llama-3.1-8B-Instruct",
  "input": "Tell me a joke."
}

Using HTTP requests:

curl https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <YOUR RUNPOD API KEY>" \
  -d '{
    "model": "<YOUR DEPLOYED MODEL REPO/NAME>",
    "input": "Tell me a joke."
  }'

Anthropic Messages API

Path: /openai/v1/messages (full URL: https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1/messages)

Supports the Anthropic Messages API format. Served directly, bypassing the RunPod queue.

{
  "model": "meta-llama/Llama-3.1-8B-Instruct",
  "max_tokens": 256,
  "messages": [
    {"role": "user", "content": "Hello!"}
  ]
}

Using HTTP requests:

curl https://api.runpod.ai/v2/<YOUR ENDPOINT ID>/openai/v1/messages \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <YOUR RUNPOD API KEY>" \
  -d '{
    "model": "<YOUR DEPLOYED MODEL REPO/NAME>",
    "max_tokens": 256,
    "messages": [
      {"role": "user", "content": "Hello!"}
    ]
  }'

Usage: Standard (Non-OpenAI)

Request Input Parameters

Click to expand table

Argument	Type	Default	Description
`prompt`	str		Prompt string to generate text based on. Proxied to `/v1/completions`.
`messages`	list[dict[str, str]]		List of messages, which will automatically have the model's chat template applied. Overrides `prompt`.
`sampling_params`	dict	{}	Sampling parameters forwarded in the request body (temperature, top_p, max_tokens, ...).
`stream`	bool	False	Whether to enable streaming of output. If True, raw SSE chunks are streamed as they are generated.

You can also call any vLLM route directly with the generic proxy form:

{
  "input": {
    "route": "/v1/chat/completions",
    "method": "POST",
    "body": {
      "model": "<YOUR DEPLOYED MODEL>",
      "messages": [{"role": "user", "content": "hello"}],
      "stream": true
    }
  }
}

Sampling Parameters

Below are all available sampling parameters that you can specify in the sampling_params dictionary. If you do not specify any of these parameters, the default values will be used.

Click to expand table

Argument	Type	Default	Description
`n`	int	1	Number of output sequences generated from the prompt. The top `n` sequences are returned.
`best_of`	Optional[int]	`n`	Number of output sequences generated from the prompt. The top `n` sequences are returned from these `best_of` sequences. Must be ≥ `n`. Treated as beam width in beam search. Default is `n`.
`presence_penalty`	float	0.0	Penalizes new tokens based on their presence in the generated text so far. Values > 0 encourage new tokens, values < 0 encourage repetition.
`frequency_penalty`	float	0.0	Penalizes new tokens based on their frequency in the generated text so far. Values > 0 encourage new tokens, values < 0 encourage repetition.
`repetition_penalty`	float	1.0	Penalizes new tokens based on their appearance in the prompt and generated text. Values > 1 encourage new tokens, values < 1 encourage repetition.
`temperature`	float	1.0	Controls the randomness of sampling. Lower values make it more deterministic, higher values make it more random. Zero means greedy sampling.
`top_p`	float	1.0	Controls the cumulative probability of top tokens to consider. Must be in (0, 1]. Set to 1 to consider all tokens.
`top_k`	int	-1	Controls the number of top tokens to consider. Set to -1 to consider all tokens.
`min_p`	float	0.0	Represents the minimum probability for a token to be considered, relative to the most likely token. Must be in [0, 1]. Set to 0 to disable.
`use_beam_search`	bool	False	Whether to use beam search instead of sampling.
`length_penalty`	float	1.0	Penalizes sequences based on their length. Used in beam search.
`early_stopping`	Union[bool, str]	False	Controls stopping condition in beam search. Can be `True`, `False`, or `"never"`.
`stop`	Union[None, str, List[str]]	None	List of strings that stop generation when produced. The output will not contain these strings.
`stop_token_ids`	Optional[List[int]]	None	List of token IDs that stop generation when produced. Output contains these tokens unless they are special tokens.
`ignore_eos`	bool	False	Whether to ignore the End-Of-Sequence token and continue generating tokens after its generation.
`max_tokens`	int	16	Maximum number of tokens to generate per output sequence.
`skip_special_tokens`	bool	True	Whether to skip special tokens in the output.
`spaces_between_special_tokens`	bool	True	Whether to add spaces between special tokens in the output.

Text Input Formats

You may either use a prompt or a list of messages as input.

prompt The prompt string can be any string, and the model's chat template will not be applied to it unless apply_chat_template is set to true, in which case it will be treated as a user message.
```
Example:
```json
{
  "input": {
    "prompt": "why sky is blue?",
    "sampling_params": {
      "temperature": 0.7,
      "max_tokens": 100
    }
  }
}
```
```

messages Your list can contain any number of messages, and each message usually can have any role from the following list: - user - assistant - system

However, some models may have different roles, so you should check the model's chat template to see which roles are required.

The model's chat template will be applied to the messages automatically, so the model must have one.

Example:

{
  "input": {
    "messages": [
      {
        "role": "system",
        "content": "You are a helpful AI assistant that provides clear and concise responses."
      },
      {
        "role": "user",
        "content": "Can you explain the difference between supervised and unsupervised learning?"
      },
      {
        "role": "assistant",
        "content": "Sure! Supervised learning uses labeled data, meaning each input has a corresponding correct output. The model learns by mapping inputs to known outputs. In contrast, unsupervised learning works with unlabeled data, where the model identifies patterns, structures, or clusters without predefined answers."
      }
    ],
    "sampling_params": {
      "temperature": 0.7,
      "max_tokens": 100
    }
  }
}

Rendered by GitHub from the current README

Read on GitHub

Top contributors

No contributor data yet — it is fetched during the contributor sync pass.

Quality score

Would you bet a product on this? Bounded 0–100 and slow moving.

83/10083

Maintenance: 90%; How recently the default branch was pushed to
Releases: 86%; Recency and cadence of tagged releases
Community: 83%; Contributor breadth and bus factor
Issue backlog: 35%; Open issues measured against the size of the audience
Documentation: 100%; README depth, plus a homepage and topics
Licensing: 100%; How freely the code can be adopted

What to know before adopting

Nothing flagged — maintained, documented and cleanly licensed.
Usable in commercial products with attribution.

Details

Created: 3 years ago
Last push: 8 days ago
Last release: last month
Release tag: v2.22.5
Releases in the last year: 30
Default branch: main
Size: 29k KB
Owner type: Organization
Owner location: United States of America
Discovered via: topic:vllm
First indexed: 13 days ago
Last synced: 12 days ago

Top AI Repos — open-source AI, indexed and scored

Trend

From the README

OpenAI-Compatible vLLM Serverless Endpoint Worker

Table of Contents

Setting up the Serverless Worker

Option 1: Deploy Any Model Using Pre-Built Docker Image [Recommended]

Configuration

Configuration File (config.yaml)

Option 2: Build Docker Image with Model Inside

Prerequisites

Arguments

Example: Building an image with OpenChat-3.5

Example: Choosing the vLLM Version

(Optional) Including Huggingface Token

Compatible Model Architectures

Usage: OpenAI Compatibility

Modifying your OpenAI Codebase to use your deployed vLLM Worker

OpenAI Request Input Parameters:

Chat Completions [RECOMMENDED]

Examples: Using your Runpod endpoint with OpenAI

Chat Completions:

Getting a list of names for available models:

OpenAI Responses API

Anthropic Messages API

Usage: Standard (Non-OpenAI)

Request Input Parameters

Sampling Parameters

Text Input Formats

Top contributors

Quality score

What to know before adopting

Categories

Details

Trend

From the README

OpenAI-Compatible vLLM Serverless Endpoint Worker

Table of Contents

Setting up the Serverless Worker

Option 1: Deploy Any Model Using Pre-Built Docker Image [Recommended]

Configuration

Configuration File (config.yaml)

Option 2: Build Docker Image with Model Inside

Prerequisites

Arguments

Example: Building an image with OpenChat-3.5

Example: Choosing the vLLM Version

(Optional) Including Huggingface Token

Compatible Model Architectures

Usage: OpenAI Compatibility

Modifying your OpenAI Codebase to use your deployed vLLM Worker

OpenAI Request Input Parameters:

Chat Completions [RECOMMENDED]

Examples: Using your Runpod endpoint with OpenAI

Chat Completions:

Getting a list of names for available models:

OpenAI Responses API

Anthropic Messages API

Usage: Standard (Non-OpenAI)

Request Input Parameters

Sampling Parameters

Text Input Formats

Top contributors

Quality score

What to know before adopting

Categories

Details