How to use Mistral Nemo chat model

Mistral Nemo chat model

Mistral Nemo is a cutting-edge Language Model (LLM) boasting state-of-the-art reasoning, world knowledge, and coding capabilities within its size category.

Mistral Nemo is a 12B model, making it a powerful drop-in replacement for any system using Mistral 7B, which it supersedes. It supports a context length of 128K, and it accepts only text inputs and generates text outputs.

Additionally, Mistral Nemo is:

• Jointly developed with Nvidia. This collaboration has resulted in a powerful 12B model that pushes the boundaries of language understanding and generation.
• Multilingual proficient. Mistral Nemo is equipped with a tokenizer called Tekken, which is designed for multilingual applications. It supports over 100 languages, such as English, French, German, and Spanish. Tekken is more efficient than the Llama 3 tokenizer in compressing text for approximately 85% of all languages, with significant improvements in Malayalam, Hindi, Arabic, and prevalent European languages.
• Agent-centric. Mistral Nemo possesses top-tier agentic capabilities, including native function calling and JSON outputting.
• Advanced in reasoning. Mistral Nemo demonstrates state-of-the-art mathematical and reasoning capabilities within its size category.

You can learn more about the models in their respective model card:

• Mistral-Nemo

Prerequisites

To use Mistral Nemo chat model with Azure AI Studio, you need the following prerequisites:

A model deployment

Deployment to serverless APIs

Mistral Nemo chat model can be deployed to serverless API endpoints with pay-as-you-go billing. This kind of deployment provides a way to consume models as an API without hosting them on your subscription, while keeping the enterprise security and compliance that organizations need.

Deployment to a serverless API endpoint doesn't require quota from your subscription. If your model isn't deployed already, use the Azure AI Studio, Azure Machine Learning SDK for Python, the Azure CLI, or ARM templates to deploy the model as a serverless API.

The inference package installed

You can consume predictions from this model by using the azure-ai-inference package with Python. To install this package, you need the following prerequisites:

• Python 3.8 or later installed, including pip.
• The endpoint URL. To construct the client library, you need to pass in the endpoint URL. The endpoint URL has the form https://your-host-name.your-azure-region.inference.ai.azure.com, where your-host-name is your unique model deployment host name and your-azure-region is the Azure region where the model is deployed (for example, eastus2).
• Depending on your model deployment and authentication preference, you need either a key to authenticate against the service, or Microsoft Entra ID credentials. The key is a 32-character string.

Once you have these prerequisites, install the Azure AI inference package with the following command:

pip install azure-ai-inference

Read more about the Azure AI inference package and reference.

Work with chat completions

In this section, you use the Azure AI model inference API with a chat completions model for chat.

Create a client to consume the model

First, create the client to consume the model. The following code uses an endpoint URL and key that are stored in environment variables.

import os
from azure.ai.inference import ChatCompletionsClient
from azure.core.credentials import AzureKeyCredential

client = ChatCompletionsClient(
    endpoint=os.environ["AZURE_INFERENCE_ENDPOINT"],
    credential=AzureKeyCredential(os.environ["AZURE_INFERENCE_CREDENTIAL"]),
)

Get the model's capabilities

The /info route returns information about the model that is deployed to the endpoint. Return the model's information by calling the following method:

model_info = client.get_model_info()

The response is as follows:

print("Model name:", model_info.model_name)
print("Model type:", model_info.model_type)
print("Model provider name:", model_info.model_provider_name)

Model name: Mistral-Nemo
Model type: chat-completions
Model provider name: MistralAI

Create a chat completion request

The following example shows how you can create a basic chat completions request to the model.

from azure.ai.inference.models import SystemMessage, UserMessage

response = client.complete(
    messages=[
        SystemMessage(content="You are a helpful assistant."),
        UserMessage(content="How many languages are in the world?"),
    ],
)

The response is as follows, where you can see the model's usage statistics:

print("Response:", response.choices[0].message.content)
print("Model:", response.model)
print("Usage:")
print("\tPrompt tokens:", response.usage.prompt_tokens)
print("\tTotal tokens:", response.usage.total_tokens)
print("\tCompletion tokens:", response.usage.completion_tokens)

Response: As of now, it's estimated that there are about 7,000 languages spoken around the world. However, this number can vary as some languages become extinct and new ones develop. It's also important to note that the number of speakers can greatly vary between languages, with some having millions of speakers and others only a few hundred.
Model: Mistral-Nemo
Usage: 
  Prompt tokens: 19
  Total tokens: 91
  Completion tokens: 72

Inspect the usage section in the response to see the number of tokens used for the prompt, the total number of tokens generated, and the number of tokens used for the completion.

Stream content

By default, the completions API returns the entire generated content in a single response. If you're generating long completions, waiting for the response can take many seconds.

You can stream the content to get it as it's being generated. Streaming content allows you to start processing the completion as content becomes available. This mode returns an object that streams back the response as data-only server-sent events. Extract chunks from the delta field, rather than the message field.

result = client.complete(
    messages=[
        SystemMessage(content="You are a helpful assistant."),
        UserMessage(content="How many languages are in the world?"),
    ],
    temperature=0,
    top_p=1,
    max_tokens=2048,
    stream=True,
)

To stream completions, set stream=True when you call the model.

To visualize the output, define a helper function to print the stream.

def print_stream(result):
    """
    Prints the chat completion with streaming.
    """
    import time
    for update in result:
        if update.choices:
            print(update.choices[0].delta.content, end="")

You can visualize how streaming generates content:

print_stream(result)

Explore more parameters supported by the inference client

Explore other parameters that you can specify in the inference client. For a full list of all the supported parameters and their corresponding documentation, see Azure AI Model Inference API reference.

from azure.ai.inference.models import ChatCompletionsResponseFormatText

response = client.complete(
    messages=[
        SystemMessage(content="You are a helpful assistant."),
        UserMessage(content="How many languages are in the world?"),
    ],
    presence_penalty=0.1,
    frequency_penalty=0.8,
    max_tokens=2048,
    stop=["<|endoftext|>"],
    temperature=0,
    top_p=1,
    response_format={ "type": ChatCompletionsResponseFormatText() },
)

If you want to pass a parameter that isn't in the list of supported parameters, you can pass it to the underlying model using extra parameters. See Pass extra parameters to the model.

Create JSON outputs

Mistral Nemo chat model can create JSON outputs. Set response_format to json_object to enable JSON mode and guarantee that the message the model generates is valid JSON. You must also instruct the model to produce JSON yourself via a system or user message. Also, the message content might be partially cut off if finish_reason="length", which indicates that the generation exceeded max_tokens or that the conversation exceeded the max context length.

from azure.ai.inference.models import ChatCompletionsResponseFormatJSON

response = client.complete(
    messages=[
        SystemMessage(content="You are a helpful assistant that always generate responses in JSON format, using."
                      " the following format: { ""answer"": ""response"" }."),
        UserMessage(content="How many languages are in the world?"),
    ],
    response_format={ "type": ChatCompletionsResponseFormatJSON() }
)

Pass extra parameters to the model

The Azure AI Model Inference API allows you to pass extra parameters to the model. The following code example shows how to pass the extra parameter logprobs to the model.

Before you pass extra parameters to the Azure AI model inference API, make sure your model supports those extra parameters. When the request is made to the underlying model, the header extra-parameters is passed to the model with the value pass-through. This value tells the endpoint to pass the extra parameters to the model. Use of extra parameters with the model doesn't guarantee that the model can actually handle them. Read the model's documentation to understand which extra parameters are supported.

response = client.complete(
    messages=[
        SystemMessage(content="You are a helpful assistant."),
        UserMessage(content="How many languages are in the world?"),
    ],
    model_extras={
        "logprobs": True
    }
)

The following extra parameters can be passed to Mistral Nemo chat model:

Safe mode

Mistral Nemo chat model support the parameter safe_prompt. You can toggle the safe prompt to prepend your messages with the following system prompt:

Always assist with care, respect, and truth. Respond with utmost utility yet securely. Avoid harmful, unethical, prejudiced, or negative content. Ensure replies promote fairness and positivity.

The Azure AI Model Inference API allows you to pass this extra parameter as follows:

response = client.complete(
    messages=[
        SystemMessage(content="You are a helpful assistant."),
        UserMessage(content="How many languages are in the world?"),
    ],
    model_extras={
        "safe_mode": True
    }
)

Use tools

Mistral Nemo chat model support the use of tools, which can be an extraordinary resource when you need to offload specific tasks from the language model and instead rely on a more deterministic system or even a different language model. The Azure AI Model Inference API allows you to define tools in the following way.

The following code example creates a tool definition that is able to look from flight information from two different cities.

from azure.ai.inference.models import FunctionDefinition, ChatCompletionsFunctionToolDefinition

flight_info = ChatCompletionsFunctionToolDefinition(
    function=FunctionDefinition(
        name="get_flight_info",
        description="Returns information about the next flight between two cities. This includes the name of the airline, flight number and the date and time of the next flight",
        parameters={
            "type": "object",
            "properties": {
                "origin_city": {
                    "type": "string",
                    "description": "The name of the city where the flight originates",
                },
                "destination_city": {
                    "type": "string",
                    "description": "The flight destination city",
                },
            },
            "required": ["origin_city", "destination_city"],
        },
    )
)

tools = [flight_info]

In this example, the function's output is that there are no flights available for the selected route, but the user should consider taking a train.

def get_flight_info(loc_origin: str, loc_destination: str):
    return { 
        "info": f"There are no flights available from {loc_origin} to {loc_destination}. You should take a train, specially if it helps to reduce CO2 emissions."
    }

Prompt the model to book flights with the help of this function:

messages = [
    SystemMessage(
        content="You are a helpful assistant that help users to find information about traveling, how to get"
                " to places and the different transportations options. You care about the environment and you"
                " always have that in mind when answering inqueries.",
    ),
    UserMessage(
        content="When is the next flight from Miami to Seattle?",
    ),
]

response = client.complete(
    messages=messages, tools=tools, tool_choice="auto"
)

You can inspect the response to find out if a tool needs to be called. Inspect the finish reason to determine if the tool should be called. Remember that multiple tool types can be indicated. This example demonstrates a tool of type function.

response_message = response.choices[0].message
tool_calls = response_message.tool_calls

print("Finish reason:", response.choices[0].finish_reason)
print("Tool call:", tool_calls)

To continue, append this message to the chat history:

messages.append(
    response_message
)

Now, it's time to call the appropriate function to handle the tool call. The following code snippet iterates over all the tool calls indicated in the response and calls the corresponding function with the appropriate parameters. The response is also appended to the chat history.

import json
from azure.ai.inference.models import ToolMessage

for tool_call in tool_calls:

    # Get the tool details:

    function_name = tool_call.function.name
    function_args = json.loads(tool_call.function.arguments.replace("\'", "\""))
    tool_call_id = tool_call.id

    print(f"Calling function `{function_name}` with arguments {function_args}")

    # Call the function defined above using `locals()`, which returns the list of all functions 
    # available in the scope as a dictionary. Notice that this is just done as a simple way to get
    # the function callable from its string name. Then we can call it with the corresponding
    # arguments.

    callable_func = locals()[function_name]
    function_response = callable_func(**function_args)

    print("->", function_response)

    # Once we have a response from the function and its arguments, we can append a new message to the chat 
    # history. Notice how we are telling to the model that this chat message came from a tool:

    messages.append(
        ToolMessage(
            tool_call_id=tool_call_id,
            content=json.dumps(function_response)
        )
    )

View the response from the model:

response = client.complete(
    messages=messages,
    tools=tools,
)

Apply content safety

The Azure AI model inference API supports Azure AI content safety. When you use deployments with Azure AI content safety turned on, inputs and outputs pass through an ensemble of classification models aimed at detecting and preventing the output of harmful content. The content filtering (preview) system detects and takes action on specific categories of potentially harmful content in both input prompts and output completions.

The following example shows how to handle events when the model detects harmful content in the input prompt and content safety is enabled.

from azure.ai.inference.models import AssistantMessage, UserMessage, SystemMessage

try:
    response = client.complete(
        messages=[
            SystemMessage(content="You are an AI assistant that helps people find information."),
            UserMessage(content="Chopping tomatoes and cutting them into cubes or wedges are great ways to practice your knife skills."),
        ]
    )

    print(response.choices[0].message.content)

except HttpResponseError as ex:
    if ex.status_code == 400:
        response = ex.response.json()
        if isinstance(response, dict) and "error" in response:
            print(f"Your request triggered an {response['error']['code']} error:\n\t {response['error']['message']}")
        else:
            raise
    raise

Mistral Nemo chat model

Mistral Nemo is a cutting-edge Language Model (LLM) boasting state-of-the-art reasoning, world knowledge, and coding capabilities within its size category.

Mistral Nemo is a 12B model, making it a powerful drop-in replacement for any system using Mistral 7B, which it supersedes. It supports a context length of 128K, and it accepts only text inputs and generates text outputs.

Additionally, Mistral Nemo is:

• Jointly developed with Nvidia. This collaboration has resulted in a powerful 12B model that pushes the boundaries of language understanding and generation.
• Multilingual proficient. Mistral Nemo is equipped with a tokenizer called Tekken, which is designed for multilingual applications. It supports over 100 languages, such as English, French, German, and Spanish. Tekken is more efficient than the Llama 3 tokenizer in compressing text for approximately 85% of all languages, with significant improvements in Malayalam, Hindi, Arabic, and prevalent European languages.
• Agent-centric. Mistral Nemo possesses top-tier agentic capabilities, including native function calling and JSON outputting.
• Advanced in reasoning. Mistral Nemo demonstrates state-of-the-art mathematical and reasoning capabilities within its size category.

You can learn more about the models in their respective model card:

• Mistral-Nemo

Prerequisites

To use Mistral Nemo chat model with Azure AI Studio, you need the following prerequisites:

A model deployment

Deployment to serverless APIs

Mistral Nemo chat model can be deployed to serverless API endpoints with pay-as-you-go billing. This kind of deployment provides a way to consume models as an API without hosting them on your subscription, while keeping the enterprise security and compliance that organizations need.

Deployment to a serverless API endpoint doesn't require quota from your subscription. If your model isn't deployed already, use the Azure AI Studio, Azure Machine Learning SDK for Python, the Azure CLI, or ARM templates to deploy the model as a serverless API.

The inference package installed

You can consume predictions from this model by using the @azure-rest/ai-inference package from npm. To install this package, you need the following prerequisites:

• LTS versions of Node.js with npm.
• The endpoint URL. To construct the client library, you need to pass in the endpoint URL. The endpoint URL has the form https://your-host-name.your-azure-region.inference.ai.azure.com, where your-host-name is your unique model deployment host name and your-azure-region is the Azure region where the model is deployed (for example, eastus2).
• Depending on your model deployment and authentication preference, you need either a key to authenticate against the service, or Microsoft Entra ID credentials. The key is a 32-character string.

Once you have these prerequisites, install the Azure Inference library for JavaScript with the following command:

npm install @azure-rest/ai-inference

Work with chat completions

In this section, you use the Azure AI model inference API with a chat completions model for chat.

Create a client to consume the model

First, create the client to consume the model. The following code uses an endpoint URL and key that are stored in environment variables.

import ModelClient from "@azure-rest/ai-inference";
import { isUnexpected } from "@azure-rest/ai-inference";
import { AzureKeyCredential } from "@azure/core-auth";

const client = new ModelClient(
    process.env.AZURE_INFERENCE_ENDPOINT, 
    new AzureKeyCredential(process.env.AZURE_INFERENCE_CREDENTIAL)
);

Get the model's capabilities

The /info route returns information about the model that is deployed to the endpoint. Return the model's information by calling the following method:

var model_info = await client.path("/info").get()

The response is as follows:

console.log("Model name: ", model_info.body.model_name)
console.log("Model type: ", model_info.body.model_type)
console.log("Model provider name: ", model_info.body.model_provider_name)

Model name: Mistral-Nemo
Model type: chat-completions
Model provider name: MistralAI

Create a chat completion request

The following example shows how you can create a basic chat completions request to the model.

var messages = [
    { role: "system", content: "You are a helpful assistant" },
    { role: "user", content: "How many languages are in the world?" },
];

var response = await client.path("/chat/completions").post({
    body: {
        messages: messages,
    }
});

The response is as follows, where you can see the model's usage statistics:

if (isUnexpected(response)) {
    throw response.body.error;
}

console.log("Response: ", response.body.choices[0].message.content);
console.log("Model: ", response.body.model);
console.log("Usage:");
console.log("\tPrompt tokens:", response.body.usage.prompt_tokens);
console.log("\tTotal tokens:", response.body.usage.total_tokens);
console.log("\tCompletion tokens:", response.body.usage.completion_tokens);

Response: As of now, it's estimated that there are about 7,000 languages spoken around the world. However, this number can vary as some languages become extinct and new ones develop. It's also important to note that the number of speakers can greatly vary between languages, with some having millions of speakers and others only a few hundred.
Model: Mistral-Nemo
Usage: 
  Prompt tokens: 19
  Total tokens: 91
  Completion tokens: 72

Inspect the usage section in the response to see the number of tokens used for the prompt, the total number of tokens generated, and the number of tokens used for the completion.

Stream content

By default, the completions API returns the entire generated content in a single response. If you're generating long completions, waiting for the response can take many seconds.

You can stream the content to get it as it's being generated. Streaming content allows you to start processing the completion as content becomes available. This mode returns an object that streams back the response as data-only server-sent events. Extract chunks from the delta field, rather than the message field.

var messages = [
    { role: "system", content: "You are a helpful assistant" },
    { role: "user", content: "How many languages are in the world?" },
];

var response = await client.path("/chat/completions").post({
    body: {
        messages: messages,
    }
}).asNodeStream();

To stream completions, use .asNodeStream() when you call the model.

You can visualize how streaming generates content:

var stream = response.body;
if (!stream) {
    stream.destroy();
    throw new Error(`Failed to get chat completions with status: ${response.status}`);
}

if (response.status !== "200") {
    throw new Error(`Failed to get chat completions: ${response.body.error}`);
}

var sses = createSseStream(stream);

for await (const event of sses) {
    if (event.data === "[DONE]") {
        return;
    }
    for (const choice of (JSON.parse(event.data)).choices) {
        console.log(choice.delta?.content ?? "");
    }
}

Explore more parameters supported by the inference client

Explore other parameters that you can specify in the inference client. For a full list of all the supported parameters and their corresponding documentation, see Azure AI Model Inference API reference.

var messages = [
    { role: "system", content: "You are a helpful assistant" },
    { role: "user", content: "How many languages are in the world?" },
];

var response = await client.path("/chat/completions").post({
    body: {
        messages: messages,
        presence_penalty: "0.1",
        frequency_penalty: "0.8",
        max_tokens: 2048,
        stop: ["<|endoftext|>"],
        temperature: 0,
        top_p: 1,
        response_format: { type: "text" },
    }
});

If you want to pass a parameter that isn't in the list of supported parameters, you can pass it to the underlying model using extra parameters. See Pass extra parameters to the model.

Create JSON outputs

Mistral Nemo chat model can create JSON outputs. Set response_format to json_object to enable JSON mode and guarantee that the message the model generates is valid JSON. You must also instruct the model to produce JSON yourself via a system or user message. Also, the message content might be partially cut off if finish_reason="length", which indicates that the generation exceeded max_tokens or that the conversation exceeded the max context length.

var messages = [
    { role: "system", content: "You are a helpful assistant that always generate responses in JSON format, using."
        + " the following format: { \"answer\": \"response\" }." },
    { role: "user", content: "How many languages are in the world?" },
];

var response = await client.path("/chat/completions").post({
    body: {
        messages: messages,
        response_format: { type: "json_object" }
    }
});

Pass extra parameters to the model

The Azure AI Model Inference API allows you to pass extra parameters to the model. The following code example shows how to pass the extra parameter logprobs to the model.

Before you pass extra parameters to the Azure AI model inference API, make sure your model supports those extra parameters. When the request is made to the underlying model, the header extra-parameters is passed to the model with the value pass-through. This value tells the endpoint to pass the extra parameters to the model. Use of extra parameters with the model doesn't guarantee that the model can actually handle them. Read the model's documentation to understand which extra parameters are supported.

var messages = [
    { role: "system", content: "You are a helpful assistant" },
    { role: "user", content: "How many languages are in the world?" },
];

var response = await client.path("/chat/completions").post({
    headers: {
        "extra-params": "pass-through"
    },
    body: {
        messages: messages,
        logprobs: true
    }
});

The following extra parameters can be passed to Mistral Nemo chat model:

Safe mode

Mistral Nemo chat model support the parameter safe_prompt. You can toggle the safe prompt to prepend your messages with the following system prompt:

Always assist with care, respect, and truth. Respond with utmost utility yet securely. Avoid harmful, unethical, prejudiced, or negative content. Ensure replies promote fairness and positivity.

The Azure AI Model Inference API allows you to pass this extra parameter as follows:

var messages = [
    { role: "system", content: "You are a helpful assistant" },
    { role: "user", content: "How many languages are in the world?" },
];

var response = await client.path("/chat/completions").post({
    headers: {
        "extra-params": "pass-through"
    },
    body: {
        messages: messages,
        safe_mode: true
    }
});

Use tools

Mistral Nemo chat model support the use of tools, which can be an extraordinary resource when you need to offload specific tasks from the language model and instead rely on a more deterministic system or even a different language model. The Azure AI Model Inference API allows you to define tools in the following way.

The following code example creates a tool definition that is able to look from flight information from two different cities.

const flight_info = {
    name: "get_flight_info",
    description: "Returns information about the next flight between two cities. This includes the name of the airline, flight number and the date and time of the next flight",
    parameters: {
        type: "object",
        properties: {
            origin_city: {
                type: "string",
                description: "The name of the city where the flight originates",
            },
            destination_city: {
                type: "string",
                description: "The flight destination city",
            },
        },
        required: ["origin_city", "destination_city"],
    },
}

const tools = [
    {
        type: "function",
        function: flight_info,
    },
];

In this example, the function's output is that there are no flights available for the selected route, but the user should consider taking a train.

function get_flight_info(loc_origin, loc_destination) {
    return {
        info: "There are no flights available from " + loc_origin + " to " + loc_destination + ". You should take a train, specially if it helps to reduce CO2 emissions."
    }
}

Prompt the model to book flights with the help of this function:

var result = await client.path("/chat/completions").post({
    body: {
        messages: messages,
        tools: tools,
        tool_choice: "auto"
    }
});

You can inspect the response to find out if a tool needs to be called. Inspect the finish reason to determine if the tool should be called. Remember that multiple tool types can be indicated. This example demonstrates a tool of type function.

const response_message = response.body.choices[0].message;
const tool_calls = response_message.tool_calls;

console.log("Finish reason: " + response.body.choices[0].finish_reason);
console.log("Tool call: " + tool_calls);

To continue, append this message to the chat history:

messages.push(response_message);

Now, it's time to call the appropriate function to handle the tool call. The following code snippet iterates over all the tool calls indicated in the response and calls the corresponding function with the appropriate parameters. The response is also appended to the chat history.

function applyToolCall({ function: call, id }) {
    // Get the tool details:
    const tool_params = JSON.parse(call.arguments);
    console.log("Calling function " + call.name + " with arguments " + tool_params);

    // Call the function defined above using `window`, which returns the list of all functions 
    // available in the scope as a dictionary. Notice that this is just done as a simple way to get
    // the function callable from its string name. Then we can call it with the corresponding
    // arguments.
    const function_response = tool_params.map(window[call.name]);
    console.log("-> " + function_response);

    return function_response
}

for (const tool_call of tool_calls) {
    var tool_response = tool_call.apply(applyToolCall);

    messages.push(
        {
            role: "tool",
            tool_call_id: tool_call.id,
            content: tool_response
        }
    );
}

View the response from the model:

var result = await client.path("/chat/completions").post({
    body: {
        messages: messages,
        tools: tools,
    }
});

Apply content safety

The Azure AI model inference API supports Azure AI content safety. When you use deployments with Azure AI content safety turned on, inputs and outputs pass through an ensemble of classification models aimed at detecting and preventing the output of harmful content. The content filtering (preview) system detects and takes action on specific categories of potentially harmful content in both input prompts and output completions.

The following example shows how to handle events when the model detects harmful content in the input prompt and content safety is enabled.

try {
    var messages = [
        { role: "system", content: "You are an AI assistant that helps people find information." },
        { role: "user", content: "Chopping tomatoes and cutting them into cubes or wedges are great ways to practice your knife skills." },
    ];

    var response = await client.path("/chat/completions").post({
        body: {
            messages: messages,
        }
    });

    console.log(response.body.choices[0].message.content);
}
catch (error) {
    if (error.status_code == 400) {
        var response = JSON.parse(error.response._content);
        if (response.error) {
            console.log(`Your request triggered an ${response.error.code} error:\n\t ${response.error.message}`);
        }
        else
        {
            throw error;
        }
    }
}

Mistral Nemo chat model

Mistral Nemo is a cutting-edge Language Model (LLM) boasting state-of-the-art reasoning, world knowledge, and coding capabilities within its size category.

Mistral Nemo is a 12B model, making it a powerful drop-in replacement for any system using Mistral 7B, which it supersedes. It supports a context length of 128K, and it accepts only text inputs and generates text outputs.

Additionally, Mistral Nemo is:

• Jointly developed with Nvidia. This collaboration has resulted in a powerful 12B model that pushes the boundaries of language understanding and generation.
• Multilingual proficient. Mistral Nemo is equipped with a tokenizer called Tekken, which is designed for multilingual applications. It supports over 100 languages, such as English, French, German, and Spanish. Tekken is more efficient than the Llama 3 tokenizer in compressing text for approximately 85% of all languages, with significant improvements in Malayalam, Hindi, Arabic, and prevalent European languages.
• Agent-centric. Mistral Nemo possesses top-tier agentic capabilities, including native function calling and JSON outputting.
• Advanced in reasoning. Mistral Nemo demonstrates state-of-the-art mathematical and reasoning capabilities within its size category.

You can learn more about the models in their respective model card:

• Mistral-Nemo

Prerequisites

To use Mistral Nemo chat model with Azure AI Studio, you need the following prerequisites:

A model deployment

Deployment to serverless APIs

Mistral Nemo chat model can be deployed to serverless API endpoints with pay-as-you-go billing. This kind of deployment provides a way to consume models as an API without hosting them on your subscription, while keeping the enterprise security and compliance that organizations need.

Deployment to a serverless API endpoint doesn't require quota from your subscription. If your model isn't deployed already, use the Azure AI Studio, Azure Machine Learning SDK for Python, the Azure CLI, or ARM templates to deploy the model as a serverless API.

The inference package installed

You can consume predictions from this model by using the Azure.AI.Inference package from NuGet. To install this package, you need the following prerequisites:

• The endpoint URL. To construct the client library, you need to pass in the endpoint URL. The endpoint URL has the form https://your-host-name.your-azure-region.inference.ai.azure.com, where your-host-name is your unique model deployment host name and your-azure-region is the Azure region where the model is deployed (for example, eastus2).
• Depending on your model deployment and authentication preference, you need either a key to authenticate against the service, or Microsoft Entra ID credentials. The key is a 32-character string.

Once you have these prerequisites, install the Azure AI inference library with the following command:

dotnet add package Azure.AI.Inference --prerelease

You can also authenticate with Microsoft Entra ID (formerly Azure Active Directory). To use credential providers provided with the Azure SDK, install the Azure.Identity package:

dotnet add package Azure.Identity

Import the following namespaces:

using Azure;
using Azure.Identity;
using Azure.AI.Inference;

This example also uses the following namespaces but you may not always need them:

using System.Text.Json;
using System.Text.Json.Serialization;
using System.Reflection;

Work with chat completions

In this section, you use the Azure AI model inference API with a chat completions model for chat.

Create a client to consume the model

First, create the client to consume the model. The following code uses an endpoint URL and key that are stored in environment variables.

ChatCompletionsClient client = new ChatCompletionsClient(
    new Uri(Environment.GetEnvironmentVariable("AZURE_INFERENCE_ENDPOINT")),
    new AzureKeyCredential(Environment.GetEnvironmentVariable("AZURE_INFERENCE_CREDENTIAL"))
);

Get the model's capabilities

The /info route returns information about the model that is deployed to the endpoint. Return the model's information by calling the following method:

Response<ModelInfo> modelInfo = client.GetModelInfo();

The response is as follows:

Console.WriteLine($"Model name: {modelInfo.Value.ModelName}");
Console.WriteLine($"Model type: {modelInfo.Value.ModelType}");
Console.WriteLine($"Model provider name: {modelInfo.Value.ModelProviderName}");

Model name: Mistral-Nemo
Model type: chat-completions
Model provider name: MistralAI

Create a chat completion request

The following example shows how you can create a basic chat completions request to the model.

ChatCompletionsOptions requestOptions = new ChatCompletionsOptions()
{
    Messages = {
        new ChatRequestSystemMessage("You are a helpful assistant."),
        new ChatRequestUserMessage("How many languages are in the world?")
    },
};

Response<ChatCompletions> response = client.Complete(requestOptions);

The response is as follows, where you can see the model's usage statistics:

Console.WriteLine($"Response: {response.Value.Choices[0].Message.Content}");
Console.WriteLine($"Model: {response.Value.Model}");
Console.WriteLine("Usage:");
Console.WriteLine($"\tPrompt tokens: {response.Value.Usage.PromptTokens}");
Console.WriteLine($"\tTotal tokens: {response.Value.Usage.TotalTokens}");
Console.WriteLine($"\tCompletion tokens: {response.Value.Usage.CompletionTokens}");

Response: As of now, it's estimated that there are about 7,000 languages spoken around the world. However, this number can vary as some languages become extinct and new ones develop. It's also important to note that the number of speakers can greatly vary between languages, with some having millions of speakers and others only a few hundred.
Model: Mistral-Nemo
Usage: 
  Prompt tokens: 19
  Total tokens: 91
  Completion tokens: 72

Inspect the usage section in the response to see the number of tokens used for the prompt, the total number of tokens generated, and the number of tokens used for the completion.

Stream content

By default, the completions API returns the entire generated content in a single response. If you're generating long completions, waiting for the response can take many seconds.

You can stream the content to get it as it's being generated. Streaming content allows you to start processing the completion as content becomes available. This mode returns an object that streams back the response as data-only server-sent events. Extract chunks from the delta field, rather than the message field.

static async Task StreamMessageAsync(ChatCompletionsClient client)
{
    ChatCompletionsOptions requestOptions = new ChatCompletionsOptions()
    {
        Messages = {
            new ChatRequestSystemMessage("You are a helpful assistant."),
            new ChatRequestUserMessage("How many languages are in the world? Write an essay about it.")
        },
        MaxTokens=4096
    };

    StreamingResponse<StreamingChatCompletionsUpdate> streamResponse = await client.CompleteStreamingAsync(requestOptions);

    await PrintStream(streamResponse);
}

To stream completions, use CompleteStreamingAsync method when you call the model. Notice that in this example we the call is wrapped in an asynchronous method.

To visualize the output, define an asynchronous method to print the stream in the console.

static async Task PrintStream(StreamingResponse<StreamingChatCompletionsUpdate> response)
{
    await foreach (StreamingChatCompletionsUpdate chatUpdate in response)
    {
        if (chatUpdate.Role.HasValue)
        {
            Console.Write($"{chatUpdate.Role.Value.ToString().ToUpperInvariant()}: ");
        }
        if (!string.IsNullOrEmpty(chatUpdate.ContentUpdate))
        {
            Console.Write(chatUpdate.ContentUpdate);
        }
    }
}

You can visualize how streaming generates content:

StreamMessageAsync(client).GetAwaiter().GetResult();

Explore more parameters supported by the inference client

Explore other parameters that you can specify in the inference client. For a full list of all the supported parameters and their corresponding documentation, see Azure AI Model Inference API reference.

requestOptions = new ChatCompletionsOptions()
{
    Messages = {
        new ChatRequestSystemMessage("You are a helpful assistant."),
        new ChatRequestUserMessage("How many languages are in the world?")
    },
    PresencePenalty = 0.1f,
    FrequencyPenalty = 0.8f,
    MaxTokens = 2048,
    StopSequences = { "<|endoftext|>" },
    Temperature = 0,
    NucleusSamplingFactor = 1,
    ResponseFormat = new ChatCompletionsResponseFormatText()
};

response = client.Complete(requestOptions);
Console.WriteLine($"Response: {response.Value.Choices[0].Message.Content}");

If you want to pass a parameter that isn't in the list of supported parameters, you can pass it to the underlying model using extra parameters. See Pass extra parameters to the model.

Create JSON outputs

Mistral Nemo chat model can create JSON outputs. Set response_format to json_object to enable JSON mode and guarantee that the message the model generates is valid JSON. You must also instruct the model to produce JSON yourself via a system or user message. Also, the message content might be partially cut off if finish_reason="length", which indicates that the generation exceeded max_tokens or that the conversation exceeded the max context length.

requestOptions = new ChatCompletionsOptions()
{
    Messages = {
        new ChatRequestSystemMessage(
            "You are a helpful assistant that always generate responses in JSON format, " +
            "using. the following format: { \"answer\": \"response\" }."
        ),
        new ChatRequestUserMessage(
            "How many languages are in the world?"
        )
    },
    ResponseFormat = new ChatCompletionsResponseFormatJSON()
};

response = client.Complete(requestOptions);
Console.WriteLine($"Response: {response.Value.Choices[0].Message.Content}");

Pass extra parameters to the model

The Azure AI Model Inference API allows you to pass extra parameters to the model. The following code example shows how to pass the extra parameter logprobs to the model.

Before you pass extra parameters to the Azure AI model inference API, make sure your model supports those extra parameters. When the request is made to the underlying model, the header extra-parameters is passed to the model with the value pass-through. This value tells the endpoint to pass the extra parameters to the model. Use of extra parameters with the model doesn't guarantee that the model can actually handle them. Read the model's documentation to understand which extra parameters are supported.

requestOptions = new ChatCompletionsOptions()
{
    Messages = {
        new ChatRequestSystemMessage("You are a helpful assistant."),
        new ChatRequestUserMessage("How many languages are in the world?")
    },
    AdditionalProperties = { { "logprobs", BinaryData.FromString("true") } },
};

response = client.Complete(requestOptions, extraParams: ExtraParameters.PassThrough);
Console.WriteLine($"Response: {response.Value.Choices[0].Message.Content}");

The following extra parameters can be passed to Mistral Nemo chat model:

Safe mode

Mistral Nemo chat model support the parameter safe_prompt. You can toggle the safe prompt to prepend your messages with the following system prompt:

Always assist with care, respect, and truth. Respond with utmost utility yet securely. Avoid harmful, unethical, prejudiced, or negative content. Ensure replies promote fairness and positivity.

The Azure AI Model Inference API allows you to pass this extra parameter as follows:

requestOptions = new ChatCompletionsOptions()
{
    Messages = {
        new ChatRequestSystemMessage("You are a helpful assistant."),
        new ChatRequestUserMessage("How many languages are in the world?")
    },
    AdditionalProperties = { { "safe_mode", BinaryData.FromString("true") } },
};

response = client.Complete(requestOptions, extraParams: ExtraParameters.PassThrough);
Console.WriteLine($"Response: {response.Value.Choices[0].Message.Content}");

Use tools

Mistral Nemo chat model support the use of tools, which can be an extraordinary resource when you need to offload specific tasks from the language model and instead rely on a more deterministic system or even a different language model. The Azure AI Model Inference API allows you to define tools in the following way.

The following code example creates a tool definition that is able to look from flight information from two different cities.

FunctionDefinition flightInfoFunction = new FunctionDefinition("getFlightInfo")
{
    Description = "Returns information about the next flight between two cities. This includes the name of the airline, flight number and the date and time of the next flight",
    Parameters = BinaryData.FromObjectAsJson(new
    {
        Type = "object",
        Properties = new
        {
            origin_city = new
            {
                Type = "string",
                Description = "The name of the city where the flight originates"
            },
            destination_city = new
            {
                Type = "string",
                Description = "The flight destination city"
            }
        }
    },
        new JsonSerializerOptions() { PropertyNamingPolicy = JsonNamingPolicy.CamelCase }
    )
};

ChatCompletionsFunctionToolDefinition getFlightTool = new ChatCompletionsFunctionToolDefinition(flightInfoFunction);

In this example, the function's output is that there are no flights available for the selected route, but the user should consider taking a train.

static string getFlightInfo(string loc_origin, string loc_destination)
{
    return JsonSerializer.Serialize(new
    {
        info = $"There are no flights available from {loc_origin} to {loc_destination}. You " +
        "should take a train, specially if it helps to reduce CO2 emissions."
    });
}

Prompt the model to book flights with the help of this function:

var chatHistory = new List<ChatRequestMessage>(){
        new ChatRequestSystemMessage(
            "You are a helpful assistant that help users to find information about traveling, " +
            "how to get to places and the different transportations options. You care about the" +
            "environment and you always have that in mind when answering inqueries."
        ),
        new ChatRequestUserMessage("When is the next flight from Miami to Seattle?")
    };

requestOptions = new ChatCompletionsOptions(chatHistory);
requestOptions.Tools.Add(getFlightTool);
requestOptions.ToolChoice = ChatCompletionsToolChoice.Auto;

response = client.Complete(requestOptions);

You can inspect the response to find out if a tool needs to be called. Inspect the finish reason to determine if the tool should be called. Remember that multiple tool types can be indicated. This example demonstrates a tool of type function.

var responseMenssage = response.Value.Choices[0].Message;
var toolsCall = responseMenssage.ToolCalls;

Console.WriteLine($"Finish reason: {response.Value.Choices[0].FinishReason}");
Console.WriteLine($"Tool call: {toolsCall[0].Id}");

To continue, append this message to the chat history:

requestOptions.Messages.Add(new ChatRequestAssistantMessage(response.Value.Choices[0].Message));

Now, it's time to call the appropriate function to handle the tool call. The following code snippet iterates over all the tool calls indicated in the response and calls the corresponding function with the appropriate parameters. The response is also appended to the chat history.

foreach (ChatCompletionsToolCall tool in toolsCall)
{
    if (tool is ChatCompletionsFunctionToolCall functionTool)
    {
        // Get the tool details:
        string callId = functionTool.Id;
        string toolName = functionTool.Name;
        string toolArgumentsString = functionTool.Arguments;
        Dictionary<string, object> toolArguments = JsonSerializer.Deserialize<Dictionary<string, object>>(toolArgumentsString);

        // Here you have to call the function defined. In this particular example we use 
        // reflection to find the method we definied before in an static class called 
        // `ChatCompletionsExamples`. Using reflection allows us to call a function 
        // by string name. Notice that this is just done for demonstration purposes as a 
        // simple way to get the function callable from its string name. Then we can call 
        // it with the corresponding arguments.

        var flags = BindingFlags.Instance | BindingFlags.Public | BindingFlags.NonPublic | BindingFlags.Static;
        string toolResponse = (string)typeof(ChatCompletionsExamples).GetMethod(toolName, flags).Invoke(null, toolArguments.Values.Cast<object>().ToArray());

        Console.WriteLine("->", toolResponse);
        requestOptions.Messages.Add(new ChatRequestToolMessage(toolResponse, callId));
    }
    else
        throw new Exception("Unsupported tool type");
}

View the response from the model:

response = client.Complete(requestOptions);

Apply content safety

The Azure AI model inference API supports Azure AI content safety. When you use deployments with Azure AI content safety turned on, inputs and outputs pass through an ensemble of classification models aimed at detecting and preventing the output of harmful content. The content filtering (preview) system detects and takes action on specific categories of potentially harmful content in both input prompts and output completions.

The following example shows how to handle events when the model detects harmful content in the input prompt and content safety is enabled.

try
{
    requestOptions = new ChatCompletionsOptions()
    {
        Messages = {
            new ChatRequestSystemMessage("You are an AI assistant that helps people find information."),
            new ChatRequestUserMessage(
                "Chopping tomatoes and cutting them into cubes or wedges are great ways to practice your knife skills."
            ),
        },
    };

    response = client.Complete(requestOptions);
    Console.WriteLine(response.Value.Choices[0].Message.Content);
}
catch (RequestFailedException ex)
{
    if (ex.ErrorCode == "content_filter")
    {
        Console.WriteLine($"Your query has trigger Azure Content Safety: {ex.Message}");
    }
    else
    {
        throw;
    }
}

Mistral Nemo chat model

Mistral Nemo is a cutting-edge Language Model (LLM) boasting state-of-the-art reasoning, world knowledge, and coding capabilities within its size category.

Mistral Nemo is a 12B model, making it a powerful drop-in replacement for any system using Mistral 7B, which it supersedes. It supports a context length of 128K, and it accepts only text inputs and generates text outputs.

Additionally, Mistral Nemo is:

• Jointly developed with Nvidia. This collaboration has resulted in a powerful 12B model that pushes the boundaries of language understanding and generation.
• Multilingual proficient. Mistral Nemo is equipped with a tokenizer called Tekken, which is designed for multilingual applications. It supports over 100 languages, such as English, French, German, and Spanish. Tekken is more efficient than the Llama 3 tokenizer in compressing text for approximately 85% of all languages, with significant improvements in Malayalam, Hindi, Arabic, and prevalent European languages.
• Agent-centric. Mistral Nemo possesses top-tier agentic capabilities, including native function calling and JSON outputting.
• Advanced in reasoning. Mistral Nemo demonstrates state-of-the-art mathematical and reasoning capabilities within its size category.

You can learn more about the models in their respective model card:

• Mistral-Nemo

Prerequisites

To use Mistral Nemo chat model with Azure AI Studio, you need the following prerequisites:

A model deployment

Deployment to serverless APIs

Mistral Nemo chat model can be deployed to serverless API endpoints with pay-as-you-go billing. This kind of deployment provides a way to consume models as an API without hosting them on your subscription, while keeping the enterprise security and compliance that organizations need.

Deployment to a serverless API endpoint doesn't require quota from your subscription. If your model isn't deployed already, use the Azure AI Studio, Azure Machine Learning SDK for Python, the Azure CLI, or ARM templates to deploy the model as a serverless API.

A REST client

Models deployed with the Azure AI model inference API can be consumed using any REST client. To use the REST client, you need the following prerequisites:

• To construct the requests, you need to pass in the endpoint URL. The endpoint URL has the form https://your-host-name.your-azure-region.inference.ai.azure.com, where your-host-name`` is your unique model deployment host name and your-azure-region`` is the Azure region where the model is deployed (for example, eastus2).
• Depending on your model deployment and authentication preference, you need either a key to authenticate against the service, or Microsoft Entra ID credentials. The key is a 32-character string.

Work with chat completions

In this section, you use the Azure AI model inference API with a chat completions model for chat.

Create a client to consume the model

First, create the client to consume the model. The following code uses an endpoint URL and key that are stored in environment variables.

Get the model's capabilities

The /info route returns information about the model that is deployed to the endpoint. Return the model's information by calling the following method:

GET /info HTTP/1.1
Host: <ENDPOINT_URI>
Authorization: Bearer <TOKEN>
Content-Type: application/json

The response is as follows:

{
    "model_name": "Mistral-Nemo",
    "model_type": "chat-completions",
    "model_provider_name": "MistralAI"
}

Create a chat completion request

The following example shows how you can create a basic chat completions request to the model.

{
    "messages": [
        {
            "role": "system",
            "content": "You are a helpful assistant."
        },
        {
            "role": "user",
            "content": "How many languages are in the world?"
        }
    ]
}

The response is as follows, where you can see the model's usage statistics:

{
    "id": "0a1234b5de6789f01gh2i345j6789klm",
    "object": "chat.completion",
    "created": 1718726686,
    "model": "Mistral-Nemo",
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "assistant",
                "content": "As of now, it's estimated that there are about 7,000 languages spoken around the world. However, this number can vary as some languages become extinct and new ones develop. It's also important to note that the number of speakers can greatly vary between languages, with some having millions of speakers and others only a few hundred.",
                "tool_calls": null
            },
            "finish_reason": "stop",
            "logprobs": null
        }
    ],
    "usage": {
        "prompt_tokens": 19,
        "total_tokens": 91,
        "completion_tokens": 72
    }
}

Inspect the usage section in the response to see the number of tokens used for the prompt, the total number of tokens generated, and the number of tokens used for the completion.

Stream content

By default, the completions API returns the entire generated content in a single response. If you're generating long completions, waiting for the response can take many seconds.

You can stream the content to get it as it's being generated. Streaming content allows you to start processing the completion as content becomes available. This mode returns an object that streams back the response as data-only server-sent events. Extract chunks from the delta field, rather than the message field.

{
    "messages": [
        {
            "role": "system",
            "content": "You are a helpful assistant."
        },
        {
            "role": "user",
            "content": "How many languages are in the world?"
        }
    ],
    "stream": true,
    "temperature": 0,
    "top_p": 1,
    "max_tokens": 2048
}

You can visualize how streaming generates content:

{
    "id": "23b54589eba14564ad8a2e6978775a39",
    "object": "chat.completion.chunk",
    "created": 1718726371,
    "model": "Mistral-Nemo",
    "choices": [
        {
            "index": 0,
            "delta": {
                "role": "assistant",
                "content": ""
            },
            "finish_reason": null,
            "logprobs": null
        }
    ]
}

The last message in the stream has finish_reason set, indicating the reason for the generation process to stop.

{
    "id": "23b54589eba14564ad8a2e6978775a39",
    "object": "chat.completion.chunk",
    "created": 1718726371,
    "model": "Mistral-Nemo",
    "choices": [
        {
            "index": 0,
            "delta": {
                "content": ""
            },
            "finish_reason": "stop",
            "logprobs": null
        }
    ],
    "usage": {
        "prompt_tokens": 19,
        "total_tokens": 91,
        "completion_tokens": 72
    }
}

Explore more parameters supported by the inference client

Explore other parameters that you can specify in the inference client. For a full list of all the supported parameters and their corresponding documentation, see Azure AI Model Inference API reference.

{
    "messages": [
        {
            "role": "system",
            "content": "You are a helpful assistant."
        },
        {
            "role": "user",
            "content": "How many languages are in the world?"
        }
    ],
    "presence_penalty": 0.1,
    "frequency_penalty": 0.8,
    "max_tokens": 2048,
    "stop": ["<|endoftext|>"],
    "temperature" :0,
    "top_p": 1,
    "response_format": { "type": "text" }
}

{
    "id": "0a1234b5de6789f01gh2i345j6789klm",
    "object": "chat.completion",
    "created": 1718726686,
    "model": "Mistral-Nemo",
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "assistant",
                "content": "As of now, it's estimated that there are about 7,000 languages spoken around the world. However, this number can vary as some languages become extinct and new ones develop. It's also important to note that the number of speakers can greatly vary between languages, with some having millions of speakers and others only a few hundred.",
                "tool_calls": null
            },
            "finish_reason": "stop",
            "logprobs": null
        }
    ],
    "usage": {
        "prompt_tokens": 19,
        "total_tokens": 91,
        "completion_tokens": 72
    }
}

If you want to pass a parameter that isn't in the list of supported parameters, you can pass it to the underlying model using extra parameters. See Pass extra parameters to the model.

Create JSON outputs

Mistral Nemo chat model can create JSON outputs. Set response_format to json_object to enable JSON mode and guarantee that the message the model generates is valid JSON. You must also instruct the model to produce JSON yourself via a system or user message. Also, the message content might be partially cut off if finish_reason="length", which indicates that the generation exceeded max_tokens or that the conversation exceeded the max context length.

{
    "messages": [
        {
            "role": "system",
            "content": "You are a helpful assistant that always generate responses in JSON format, using the following format: { \"answer\": \"response\" }"
        },
        {
            "role": "user",
            "content": "How many languages are in the world?"
        }
    ],
    "response_format": { "type": "json_object" }
}

{
    "id": "0a1234b5de6789f01gh2i345j6789klm",
    "object": "chat.completion",
    "created": 1718727522,
    "model": "Mistral-Nemo",
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "assistant",
                "content": "{\"answer\": \"There are approximately 7,117 living languages in the world today, according to the latest estimates. However, this number can vary as some languages become extinct and others are newly discovered or classified.\"}",
                "tool_calls": null
            },
            "finish_reason": "stop",
            "logprobs": null
        }
    ],
    "usage": {
        "prompt_tokens": 39,
        "total_tokens": 87,
        "completion_tokens": 48
    }
}

Pass extra parameters to the model

The Azure AI Model Inference API allows you to pass extra parameters to the model. The following code example shows how to pass the extra parameter logprobs to the model.

Before you pass extra parameters to the Azure AI model inference API, make sure your model supports those extra parameters. When the request is made to the underlying model, the header extra-parameters is passed to the model with the value pass-through. This value tells the endpoint to pass the extra parameters to the model. Use of extra parameters with the model doesn't guarantee that the model can actually handle them. Read the model's documentation to understand which extra parameters are supported.

POST /chat/completions HTTP/1.1
Host: <ENDPOINT_URI>
Authorization: Bearer <TOKEN>
Content-Type: application/json
extra-parameters: pass-through

{
    "messages": [
        {
            "role": "system",
            "content": "You are a helpful assistant."
        },
        {
            "role": "user",
            "content": "How many languages are in the world?"
        }
    ],
    "logprobs": true
}

The following extra parameters can be passed to Mistral Nemo chat model:

Safe mode

Mistral Nemo chat model support the parameter safe_prompt. You can toggle the safe prompt to prepend your messages with the following system prompt:

Always assist with care, respect, and truth. Respond with utmost utility yet securely. Avoid harmful, unethical, prejudiced, or negative content. Ensure replies promote fairness and positivity.

The Azure AI Model Inference API allows you to pass this extra parameter as follows:

POST /chat/completions HTTP/1.1
Host: <ENDPOINT_URI>
Authorization: Bearer <TOKEN>
Content-Type: application/json
extra-parameters: pass-through

{
    "messages": [
        {
            "role": "system",
            "content": "You are a helpful assistant."
        },
        {
            "role": "user",
            "content": "How many languages are in the world?"
        }
    ],
    "safemode": true
}

Use tools

Mistral Nemo chat model support the use of tools, which can be an extraordinary resource when you need to offload specific tasks from the language model and instead rely on a more deterministic system or even a different language model. The Azure AI Model Inference API allows you to define tools in the following way.

The following code example creates a tool definition that is able to look from flight information from two different cities.

{
    "type": "function",
    "function": {
        "name": "get_flight_info",
        "description": "Returns information about the next flight between two cities. This includes the name of the airline, flight number and the date and time of the next flight",
        "parameters": {
            "type": "object",
            "properties": {
                "origin_city": {
                    "type": "string",
                    "description": "The name of the city where the flight originates"
                },
                "destination_city": {
                    "type": "string",
                    "description": "The flight destination city"
                }
            },
            "required": [
                "origin_city",
                "destination_city"
            ]
        }
    }
}

In this example, the function's output is that there are no flights available for the selected route, but the user should consider taking a train.

Prompt the model to book flights with the help of this function:

{
    "messages": [
        {
            "role": "system",
            "content": "You are a helpful assistant that help users to find information about traveling, how to get to places and the different transportations options. You care about the environment and you always have that in mind when answering inqueries"
        },
        {
            "role": "user",
            "content": "When is the next flight from Miami to Seattle?"
        }
    ],
    "tool_choice": "auto",
    "tools": [
        {
            "type": "function",
            "function": {
                "name": "get_flight_info",
                "description": "Returns information about the next flight between two cities. This includes the name of the airline, flight number and the date and time of the next flight",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "origin_city": {
                            "type": "string",
                            "description": "The name of the city where the flight originates"
                        },
                        "destination_city": {
                            "type": "string",
                            "description": "The flight destination city"
                        }
                    },
                    "required": [
                        "origin_city",
                        "destination_city"
                    ]
                }
            }
        }
    ]
}

You can inspect the response to find out if a tool needs to be called. Inspect the finish reason to determine if the tool should be called. Remember that multiple tool types can be indicated. This example demonstrates a tool of type function.

{
    "id": "0a1234b5de6789f01gh2i345j6789klm",
    "object": "chat.completion",
    "created": 1718726007,
    "model": "Mistral-Nemo",
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "assistant",
                "content": "",
                "tool_calls": [
                    {
                        "id": "abc0dF1gh",
                        "type": "function",
                        "function": {
                            "name": "get_flight_info",
                            "arguments": "{\"origin_city\": \"Miami\", \"destination_city\": \"Seattle\"}",
                            "call_id": null
                        }
                    }
                ]
            },
            "finish_reason": "tool_calls",
            "logprobs": null
        }
    ],
    "usage": {
        "prompt_tokens": 190,
        "total_tokens": 226,
        "completion_tokens": 36
    }
}

To continue, append this message to the chat history:

Now, it's time to call the appropriate function to handle the tool call. The following code snippet iterates over all the tool calls indicated in the response and calls the corresponding function with the appropriate parameters. The response is also appended to the chat history.

View the response from the model:

{
    "messages": [
        {
            "role": "system",
            "content": "You are a helpful assistant that help users to find information about traveling, how to get to places and the different transportations options. You care about the environment and you always have that in mind when answering inqueries"
        },
        {
            "role": "user",
            "content": "When is the next flight from Miami to Seattle?"
        },
        {
            "role": "assistant",
            "content": "",
            "tool_calls": [
                {
                    "id": "abc0DeFgH",
                    "type": "function",
                    "function": {
                        "name": "get_flight_info",
                        "arguments": "{\"origin_city\": \"Miami\", \"destination_city\": \"Seattle\"}",
                        "call_id": null
                    }
                }
            ]
        },
        {
            "role": "tool",
            "content": "{ \"info\": \"There are no flights available from Miami to Seattle. You should take a train, specially if it helps to reduce CO2 emissions.\" }",
            "tool_call_id": "abc0DeFgH" 
        }
    ],
    "tool_choice": "auto",
    "tools": [
        {
            "type": "function",
            "function": {
            "name": "get_flight_info",
            "description": "Returns information about the next flight between two cities. This includes the name of the airline, flight number and the date and time of the next flight",
            "parameters":{
                "type": "object",
                "properties": {
                    "origin_city": {
                        "type": "string",
                        "description": "The name of the city where the flight originates"
                    },
                    "destination_city": {
                        "type": "string",
                        "description": "The flight destination city"
                    }
                },
                "required": ["origin_city", "destination_city"]
            }
            }
        }
    ]
}

Apply content safety

The Azure AI model inference API supports Azure AI content safety. When you use deployments with Azure AI content safety turned on, inputs and outputs pass through an ensemble of classification models aimed at detecting and preventing the output of harmful content. The content filtering (preview) system detects and takes action on specific categories of potentially harmful content in both input prompts and output completions.

The following example shows how to handle events when the model detects harmful content in the input prompt and content safety is enabled.

{
    "messages": [
        {
            "role": "system",
            "content": "You are an AI assistant that helps people find information."
        },
                {
            "role": "user",
            "content": "Chopping tomatoes and cutting them into cubes or wedges are great ways to practice your knife skills."
        }
    ]
}

{
    "error": {
        "message": "The response was filtered due to the prompt triggering Microsoft's content management policy. Please modify your prompt and retry.",
        "type": null,
        "param": "prompt",
        "code": "content_filter",
        "status": 400
    }
}