Skip to main content

Gemini - Google AI Studio

Pre-requisites​

Gemini-Pro

Sample Usage​

from litellm import completion
import os

os.environ['GEMINI_API_KEY'] = ""
response = completion(
    model="gemini/gemini-pro", 
    messages=[{"role": "user", "content": "write code for saying hi from LiteLLM"}]
)

Supported OpenAI Params​

  • temperature
  • top_p
  • max_tokens
  • stream
  • tools
  • tool_choice
  • response_format
  • n
  • stop

See Updated List

Passing Gemini Specific Params​

Response schema​

LiteLLM supports sending response_schema as a param for Gemini-1.5-Pro on Google AI Studio.

Response Schema

from litellm import completion 
import json 
import os 

os.environ['GEMINI_API_KEY'] = ""

messages = [
    {
        "role": "user",
        "content": "List 5 popular cookie recipes."
    }
]

response_schema = {
        "type": "array",
        "items": {
            "type": "object",
            "properties": {
                "recipe_name": {
                    "type": "string",
                },
            },
            "required": ["recipe_name"],
        },
    }


completion(
    model="gemini/gemini-1.5-pro", 
    messages=messages, 
    response_format={"type": "json_object", "response_schema": response_schema} # 👈 KEY CHANGE
    )

print(json.loads(completion.choices[0].message.content))

Validate Schema

To validate the response_schema, set enforce_validation: true.

from litellm import completion, JSONSchemaValidationError
try: 
    completion(
    model="gemini/gemini-1.5-pro", 
    messages=messages, 
    response_format={
        "type": "json_object", 
        "response_schema": response_schema,
        "enforce_validation": true # 👈 KEY CHANGE
    }
    )
except JSONSchemaValidationError as e: 
    print("Raw Response: {}".format(e.raw_response))
    raise e

LiteLLM will validate the response against the schema, and raise a JSONSchemaValidationError if the response does not match the schema.

JSONSchemaValidationError inherits from openai.APIError

Access the raw response with e.raw_response

GenerationConfig Params​

To pass additional GenerationConfig params - e.g. topK, just pass it in the request body of the call, and LiteLLM will pass it straight through as a key-value pair in the request body.

See Gemini GenerationConfigParams

from litellm import completion 
import json 
import os 

os.environ['GEMINI_API_KEY'] = ""

messages = [
    {
        "role": "user",
        "content": "List 5 popular cookie recipes."
    }
]

completion(
    model="gemini/gemini-1.5-pro", 
    messages=messages, 
    topK=1 # 👈 KEY CHANGE
)

print(json.loads(completion.choices[0].message.content))

Validate Schema

To validate the response_schema, set enforce_validation: true.

from litellm import completion, JSONSchemaValidationError
try: 
    completion(
    model="gemini/gemini-1.5-pro", 
    messages=messages, 
    response_format={
        "type": "json_object", 
        "response_schema": response_schema,
        "enforce_validation": true # 👈 KEY CHANGE
    }
    )
except JSONSchemaValidationError as e: 
    print("Raw Response: {}".format(e.raw_response))
    raise e

Specifying Safety Settings​

In certain use-cases you may need to make calls to the models and pass safety settigns different from the defaults. To do so, simple pass the safety_settings argument to completion or acompletion. For example:

response = completion(
    model="gemini/gemini-pro", 
    messages=[{"role": "user", "content": "write code for saying hi from LiteLLM"}],
    safety_settings=[
        {
            "category": "HARM_CATEGORY_HARASSMENT",
            "threshold": "BLOCK_NONE",
        },
        {
            "category": "HARM_CATEGORY_HATE_SPEECH",
            "threshold": "BLOCK_NONE",
        },
        {
            "category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",
            "threshold": "BLOCK_NONE",
        },
        {
            "category": "HARM_CATEGORY_DANGEROUS_CONTENT",
            "threshold": "BLOCK_NONE",
        },
    ]
)

Tool Calling​

from litellm import completion
import os
# set env
os.environ["GEMINI_API_KEY"] = ".."

tools = [
    {
        "type": "function",
        "function": {
            "name": "get_current_weather",
            "description": "Get the current weather in a given location",
            "parameters": {
                "type": "object",
                "properties": {
                    "location": {
                        "type": "string",
                        "description": "The city and state, e.g. San Francisco, CA",
                    },
                    "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]},
                },
                "required": ["location"],
            },
        },
    }
]
messages = [{"role": "user", "content": "What's the weather like in Boston today?"}]

response = completion(
    model="gemini/gemini-1.5-flash",
    messages=messages,
    tools=tools,
)
# Add any assertions, here to check response args
print(response)
assert isinstance(response.choices[0].message.tool_calls[0].function.name, str)
assert isinstance(
    response.choices[0].message.tool_calls[0].function.arguments, str
)


JSON Mode​

from litellm import completion 
import json 
import os 

os.environ['GEMINI_API_KEY'] = ""

messages = [
    {
        "role": "user",
        "content": "List 5 popular cookie recipes."
    }
]



completion(
    model="gemini/gemini-1.5-pro", 
    messages=messages, 
    response_format={"type": "json_object"} # 👈 KEY CHANGE
)

print(json.loads(completion.choices[0].message.content))
# Gemini-Pro-Vision LiteLLM Supports the following image types passed in `url` - Images with direct links - https://storage.googleapis.com/github-repo/img/gemini/intro/landmark3.jpg - Image in local storage - ./localimage.jpeg

Sample Usage​

import os
import litellm
from dotenv import load_dotenv

# Load the environment variables from .env file
load_dotenv()
os.environ["GEMINI_API_KEY"] = os.getenv('GEMINI_API_KEY')

prompt = 'Describe the image in a few sentences.'
# Note: You can pass here the URL or Path of image directly.
image_url = 'https://storage.googleapis.com/github-repo/img/gemini/intro/landmark3.jpg'

# Create the messages payload according to the documentation
messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "text",
                "text": prompt
            },
            {
                "type": "image_url",
                "image_url": {"url": image_url}
            }
        ]
    }
]

# Make the API call to Gemini model
response = litellm.completion(
    model="gemini/gemini-pro-vision",
    messages=messages,
)

# Extract the response content
content = response.get('choices', [{}])[0].get('message', {}).get('content')

# Print the result
print(content)

Context Caching​

Use Google AI Studio context caching is supported by

{
    ...,
    "cache_control": {"type": "ephemeral"}
}

in your message content block.

note

Gemini Context Caching only allows 1 block of continuous messages to be cached.

The raw request to Gemini looks like this: 

curl -X POST "https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-flash-001:generateContent?key=$GOOGLE_API_KEY" \
-H 'Content-Type: application/json' \
-d '{
      "contents": [
        {
          "parts":[{
            "text": "Please summarize this transcript"
          }],
          "role": "user"
        },
      ],
      "cachedContent": "'$CACHE_NAME'"
    }'

from litellm import completion 

for _ in range(2): 
    resp = completion(
        model="gemini/gemini-1.5-pro",
        messages=[
        # System Message
            {
                "role": "system",
                "content": [
                    {
                        "type": "text",
                        "text": "Here is the full text of a complex legal agreement" * 4000,
                        "cache_control": {"type": "ephemeral"}, # 👈 KEY CHANGE
                    }
                ],
            },
            # marked for caching with the cache_control parameter, so that this checkpoint can read from the previous cache.
            {
                "role": "user",
                "content": [
                    {
                        "type": "text",
                        "text": "What are the key terms and conditions in this agreement?",
                        "cache_control": {"type": "ephemeral"},
                    }
                ],
            }]
    )

    print(resp.usage) # 👈 2nd usage block will be less, since cached tokens used

Chat Models​

tip

We support ALL Gemini models, just set model=gemini/<any-model-on-gemini> as a prefix when sending litellm requests

Model NameFunction CallRequired OS Variables
gemini-procompletion(model='gemini/gemini-pro', messages)os.environ['GEMINI_API_KEY']
gemini-1.5-pro-latestcompletion(model='gemini/gemini-1.5-pro-latest', messages)os.environ['GEMINI_API_KEY']
gemini-pro-visioncompletion(model='gemini/gemini-pro-vision', messages)os.environ['GEMINI_API_KEY']