OpenRouter

LiteLLM supports all the text / chat / vision / embedding models from OpenRouter

Usage

import os
from litellm import completion

os.environ["OPENROUTER_API_KEY"] = ""
os.environ["OPENROUTER_API_BASE"] = "" # [OPTIONAL] defaults to https://openrouter.ai/api/v1
os.environ["OR_SITE_URL"] = "" # [OPTIONAL]
os.environ["OR_APP_NAME"] = "" # [OPTIONAL]

response = completion(
            model="openrouter/google/palm-2-chat-bison",
            messages=messages,
        )

Configuration with Environment Variables

For production environments, you can dynamically configure the base_url using environment variables:

import os
from litellm import completion

# Configure with environment variables
OPENROUTER_API_KEY = os.getenv("OPENROUTER_API_KEY")
OPENROUTER_BASE_URL = os.getenv("OPENROUTER_API_BASE", "https://openrouter.ai/api/v1")

# Set environment for LiteLLM
os.environ["OPENROUTER_API_KEY"] = OPENROUTER_API_KEY
os.environ["OPENROUTER_API_BASE"] = OPENROUTER_BASE_URL

response = completion(
    model="openrouter/google/palm-2-chat-bison",
    messages=messages,
    base_url=OPENROUTER_BASE_URL  # Explicitly pass base_url for clarity
)

This approach provides better flexibility for managing configurations across different environments (dev, staging, production) and makes it easier to switch between self-hosted and cloud endpoints.

OpenRouter Completion Models

LiteLLM supports ALL OpenRouter models, send model=openrouter/<your-openrouter-model> to send it to open router. See all openrouter models here

Model Name	Function Call
openrouter/openai/gpt-3.5-turbo	completion('openrouter/openai/gpt-3.5-turbo', messages)
openrouter/openai/gpt-3.5-turbo-16k	completion('openrouter/openai/gpt-3.5-turbo-16k', messages)
openrouter/openai/gpt-4	completion('openrouter/openai/gpt-4', messages)
openrouter/openai/gpt-4-32k	completion('openrouter/openai/gpt-4-32k', messages)
openrouter/anthropic/claude-2	completion('openrouter/anthropic/claude-2', messages)
openrouter/anthropic/claude-instant-v1	completion('openrouter/anthropic/claude-instant-v1', messages)
openrouter/google/palm-2-chat-bison	completion('openrouter/google/palm-2-chat-bison', messages)
openrouter/google/palm-2-codechat-bison	completion('openrouter/google/palm-2-codechat-bison', messages)
openrouter/meta-llama/llama-2-13b-chat	completion('openrouter/meta-llama/llama-2-13b-chat', messages)
openrouter/meta-llama/llama-2-70b-chat	completion('openrouter/meta-llama/llama-2-70b-chat', messages)

Passing OpenRouter Params - transforms, models, route

Pass transforms, models, routeas arguments to litellm.completion()

import os
from litellm import completion

os.environ["OPENROUTER_API_KEY"] = ""

response = completion(
            model="openrouter/google/palm-2-chat-bison",
            messages=messages,
            transforms = [""],
            route= ""
        )

Embedding

from litellm import embedding
import os

os.environ["OPENROUTER_API_KEY"] = "your-api-key"

response = embedding(
    model="openrouter/openai/text-embedding-3-small",
    input=["good morning from litellm", "this is another item"],
)
print(response)

Image Generation

OpenRouter supports image generation through select models like Google Gemini image generation models. LiteLLM transforms standard image generation requests to OpenRouter's chat completion format.

Supported Parameters

• size: Maps to OpenRouter's aspect_ratio format

  • 1024x1024 → 1:1 (square)
  • 1536x1024 → 3:2 (landscape)
  • 1024x1536 → 2:3 (portrait)
  • 1792x1024 → 16:9 (wide landscape)
  • 1024x1792 → 9:16 (tall portrait)

• quality: Maps to OpenRouter's image_size format (Gemini models)

  • low or standard → 1K
  • medium → 2K
  • high or hd → 4K

• n: Number of images to generate

Usage

from litellm import image_generation
import os

os.environ["OPENROUTER_API_KEY"] = "your-api-key"

# Basic image generation
response = image_generation(
    model="openrouter/google/gemini-2.5-flash-image",
    prompt="A beautiful sunset over a calm ocean",
)
print(response)

Advanced Usage with Parameters

from litellm import image_generation
import os

os.environ["OPENROUTER_API_KEY"] = "your-api-key"

# Generate high-quality landscape image
response = image_generation(
    model="openrouter/google/gemini-2.5-flash-image",
    prompt="A serene mountain landscape with a lake",
    size="1536x1024",  # Landscape format
    quality="high",     # High quality (4K)
)

# Access the generated image
image_data = response.data[0]
if image_data.b64_json:
    # Base64 encoded image
    print(f"Generated base64 image: {image_data.b64_json[:50]}...")
elif image_data.url:
    # Image URL
    print(f"Generated image URL: {image_data.url}")

Using OpenRouter-Specific Parameters

You can also pass OpenRouter-specific parameters directly using image_config:

from litellm import image_generation
import os

os.environ["OPENROUTER_API_KEY"] = "your-api-key"

response = image_generation(
    model="openrouter/google/gemini-2.5-flash-image",
    prompt="A futuristic cityscape at night",
    image_config={
        "aspect_ratio": "16:9",  # OpenRouter native format
        "image_size": "4K"       # OpenRouter native format
    }
)
print(response)

Response Format

The response follows the standard LiteLLM ImageResponse format:

{
    "created": 1703658209,
    "data": [{
        "b64_json": "iVBORw0KGgoAAAANSUhEUgAA...",  # Base64 encoded image
        "url": None,
        "revised_prompt": None
    }],
    "usage": {
        "input_tokens": 10,
        "output_tokens": 1290,
        "total_tokens": 1300
    }
}

Cost Tracking

OpenRouter provides cost information in the response, which LiteLLM automatically tracks:

response = image_generation(
    model="openrouter/google/gemini-2.5-flash-image",
    prompt="A cute baby sea otter",
)

# Cost is available in the response metadata
print(f"Request cost: ${response._hidden_params['additional_headers']['llm_provider-x-litellm-response-cost']}")

Image Edit

OpenRouter supports image editing through select models like Google Gemini image models. LiteLLM routes image edit requests to OpenRouter's chat completions endpoint with the source image sent as a base64 data URL and modalities: ["image", "text"].

Supported Models

Model	Description
openrouter/google/gemini-2.5-flash-image	Gemini 2.5 Flash with image editing

See all available image models on OpenRouter's model list.

Supported Parameters

Parameter	OpenRouter Mapping	Notes
size	image_config.aspect_ratio	1024x1024 → 1:1, 1536x1024 → 3:2, 1024x1536 → 2:3, 1792x1024 → 16:9, 1024x1792 → 9:16
quality	image_config.image_size	low/standard → 1K, medium → 2K, high/hd → 4K
n	n	Number of images

note

quality=high (4K) is only supported by google/gemini-3-pro-image-preview and google/gemini-3.1-flash-image-preview. The google/gemini-2.5-flash-image model supports up to medium (2K).

Usage

from litellm import image_edit
import os

os.environ["OPENROUTER_API_KEY"] = "your-api-key"

# Basic image edit
response = image_edit(
    model="openrouter/google/gemini-2.5-flash-image",
    image=open("original_image.png", "rb"),
    prompt="Make the sky a vibrant purple sunset",
)

print(response)

Advanced Usage with Parameters

from litellm import image_edit
import os

os.environ["OPENROUTER_API_KEY"] = "your-api-key"

# Edit with size and quality parameters
response = image_edit(
    model="openrouter/google/gemini-2.5-flash-image",
    image=open("photo.png", "rb"),
    prompt="Add northern lights to the sky",
    size="1536x1024",   # Maps to aspect_ratio 3:2
    quality="high",      # Maps to image_size 4K
)

# Access the edited image
image_data = response.data[0]
if image_data.b64_json:
    import base64
    with open("edited.png", "wb") as f:
        f.write(base64.b64decode(image_data.b64_json))

Multiple Images Edit

from litellm import image_edit
import os

os.environ["OPENROUTER_API_KEY"] = "your-api-key"

response = image_edit(
    model="openrouter/google/gemini-2.5-flash-image",
    image=[
        open("scene.png", "rb"),
        open("style_reference.png", "rb"),
    ],
    prompt="Blend the reference style into the scene",
)

print(response)

Decisions (Jev)

OpenRouter serves TypeSafe's Jev decision models on its Decisions API, POST https://openrouter.ai/api/alpha/decisions, and rejects them on /chat/completions. The LiteLLM proxy reaches that API in two ways.

Through /v1/systemone and /v1/decisions

Add the model to your model_list with the openrouter/ prefix. The model then goes through the router like any other deployment, so it gets load balancing, fallbacks and per-deployment keys, and you can call it with a System One body on /v1/systemone or an OpenAI Decisions body on /v1/decisions.

model_list:
  - model_name: jev
    litellm_params:
      model: openrouter/typesafe/jev-1.14
      api_key: os.environ/OPENROUTER_API_KEY

curl -X POST 'http://0.0.0.0:4000/v1/systemone' \
-H "Authorization: Bearer $LITELLM_API_KEY" \
-H 'Content-Type: application/json' \
-d '{
  "model": "jev",
  "state": "The sky is blue on a clear day.",
  "questions": {"blue": {"type": "noul", "instructions": "Is the sky described as blue?"}}
}'

From v1.104.3, the spend is the usage.cost that OpenRouter returns. LiteLLM sends it back in the x-litellm-response-cost header and writes it to the spend logs. Earlier releases have no price for jev-1.14, and v1.105.0-rc.3 and the v1.106.0 dev builds do not read usage.cost either. See Decisions for the formats and the supported features.

Through the /openrouter pass-through

/openrouter/{path} sends the request body to OpenRouter as it is. The key it sends comes from the first OpenRouter deployment in model_list (an openrouter/ model or custom_llm_provider: openrouter) that sets use_in_pass_through: true and has a key, either its api_key or the one in its litellm_credential_name, or from OPENROUTER_API_KEY when no deployment does. The request always goes to OPENROUTER_API_BASE, never to that deployment's api_base. Use it when you already have OpenRouter Decisions request bodies and do not need routing. It is in v1.104.0 and later, and in v1.101.1 and v1.102.1 and the later patches on those two lines, but in no v1.103 release.

The path is relative to https://openrouter.ai/api, which is OPENROUTER_API_BASE without its trailing /v1. So the Decisions API is /openrouter/alpha/decisions. Do not add /api to the path, since /openrouter/api/alpha/decisions goes to https://openrouter.ai/api/api/alpha/decisions and OpenRouter answers it with a 404.

A virtual key with a models list needs typesafe/jev-1.14, the model name in the request body, in that list to call the pass-through. A model_list alias such as jev is not enough, and the call gets a 403.

curl -X POST 'http://0.0.0.0:4000/openrouter/alpha/decisions' \
-H "Authorization: Bearer $LITELLM_API_KEY" \
-H 'Content-Type: application/json' \
-d '{
  "model": "typesafe/jev-1.14",
  "state": "The sky is blue on a clear day.",
  "questions": {"blue": {"type": "noul", "instructions": "Is the sky described as blue?"}}
}'

Response

{
  "model": "typesafe/jev-1.14-20261014",
  "answers": {"blue": {"type": "noul", "noul": 0.99}},
  "usage": {"input_tokens": 282, "output_tokens": 20, "cost": 0.000011844},
  "id": "gen-dec-1792543634-CgEHGExRXfbzgBwKxiWf",
  "provider": "TypeSafe"
}

LiteLLM logs each /alpha/decisions call under the versioned model that OpenRouter returns, for example openrouter/typesafe/jev-1.14-20261014. From v1.104.3, the spend is the usage.cost in the response, and earlier releases have no price for jev-1.14. The v1.105.0 dev builds and release candidates and the v1.106.0 dev builds do not read usage.cost either. Calls to other /openrouter paths, such as /openrouter/v1/chat/completions, are forwarded but record $0 spend, so budgets do not count them.