Author(s): Naveen Krishnan Originally published on Towards AI. Last week, I shared my thoughts on phi‑4 models and their innovative multimodal approach. Today, I’m thrilled to write about GPT‑4.5 — a model that not only pushes the boundaries of conversational AI but also makes it easier for developers to integrate powerful language capabilities into their apps via Azure OpenAI and Foundry. Grab your favorite beverage ☕, settle in, and let’s explore how GPT‑4.5 is set to transform our interactions with technology! Image Source: OpenAI News | OpenAI From GPT‑4 to GPT‑4.5: A Quick Evolutionary Recap 🔍 GPT‑4 paved the way for richer, more nuanced conversations. With GPT‑4.5, OpenAI has fine‑tuned the art of understanding context and generating responses that are even more human-like. Improvements in efficiency, contextual awareness, and multimodal integration mean that whether you’re building chatbots, content generators, or analytical tools, GPT‑4.5 can handle your toughest challenges. But the real magic happens when you combine GPT‑4.5 with the robust enterprise-grade capabilities of Azure OpenAI Service — and then manage everything seamlessly using Azure AI Foundry. The result? A platform that’s both flexible and scalable for modern app development. ✨ Key Features of GPT‑4.5 💡 Enhanced Conversational Depth: GPT‑4.5 can maintain context over longer conversations, delivering responses that feel more intuitive and relevant. Improved Accuracy & Efficiency: Faster processing means you get your answers almost in real time without sacrificing quality. Humanized Output: With its refined tone and style, GPT‑4.5’s responses feel less mechanical and more like chatting with an insightful friend. Seamless Multimodal Integration: Whether you’re feeding text, images, or data from various sources, GPT‑4.5 adapts and responds with finesse. Enterprise‑Grade Integration: Through Azure OpenAI and Foundry, GPT‑4.5 becomes a part of a secure, scalable, and fully managed ecosystem ideal for production environments. Why Azure OpenAI with Foundry? 🔗 Integrating GPT‑4.5 via Azure OpenAI Service offers several advantages: Security & Compliance: Azure ensures your data is handled in compliance with industry standards (GDPR, HIPAA, etc.). Scalability: Whether you’re a startup or an enterprise, Azure’s infrastructure scales with your needs. Unified Management: Azure AI Foundry simplifies the management of models, data sources, and endpoints. Easy Integration: With robust SDKs and clear sample code, you can quickly incorporate GPT‑4.5 into your applications. In the sections below, I’ll walk you through sample code that demonstrates how to invoke GPT‑4.5 using Azure OpenAI and Foundry — across multiple languages so you can pick the one that fits your project best. Let’s get coding! 🚀 Invoking GPT‑4.5 via Azure OpenAI Using Foundry Setting the Stage: Environment & Authentication Before diving into the code, ensure you have the following prerequisites: An Azure OpenAI Service resource with GPT‑4.5 available in your subscription. Access to Azure AI Foundry, which helps manage and connect your models. Image Source: Screenshot by User Appropriate credentials (API keys or managed identities) stored securely (e.g., in environment variables or Azure Key Vault). Below, you’ll find sample code in C# (.NET) and Python. These examples assume you have set environment variables like AZURE_OPENAI_ENDPOINT, AZURE_OPENAI_API_KEY, and AZURE_OPENAI_DEPLOYMENT_NAME. Adjust these as needed! Sample Code in C# (.NET) Below is a sample console application written in C# that initializes the Azure OpenAI client, configures the Foundry connection, and sends a request to GPT‑4.5. using System;using Azure;using Azure.AI.OpenAI;using System.Collections.Generic;namespace GPT45Demo{ class Program { static void Main(string[] args) { // Load configuration from environment variables string endpoint = Environment.GetEnvironmentVariable("AZURE_OPENAI_ENDPOINT"); string apiKey = Environment.GetEnvironmentVariable("AZURE_OPENAI_API_KEY"); string deploymentName = Environment.GetEnvironmentVariable("AZURE_OPENAI_DEPLOYMENT_NAME"); // Initialize the Azure OpenAI client using Foundry integration settings OpenAIClient client = new OpenAIClient(new Uri(endpoint), new AzureKeyCredential(apiKey)); // Create a system prompt to guide GPT-4.5's responses string systemPrompt = "You are a knowledgeable assistant with deep insights on a range of topics. Please respond in a friendly and engaging manner, using emojis where appropriate. 😊"; // Build conversation history – you could extend this to include previous interactions List<ChatMessage> messages = new List<ChatMessage> { new ChatMessage(ChatRole.System, systemPrompt), new ChatMessage(ChatRole.User, "Can you show me how to invoke GPT-4.5 using Azure OpenAI with Foundry integration?") }; // Create chat completion options ChatCompletionsOptions options = new ChatCompletionsOptions { MaxTokens = 500, Temperature = 0.7f, // Setting the deployment name from environment variables ensures we are using our GPT-4.5 model DeploymentName = deploymentName }; // Add our conversation messages foreach (var msg in messages) { options.Messages.Add(msg); } // Send the request and receive the response ChatCompletions response = client.GetChatCompletions(options); // Print the first completion result Console.WriteLine("Response from GPT-4.5:"); Console.WriteLine(response.Choices[0].Message.Content); } }} Explanation: We begin by loading our endpoint, API key, and deployment name from environment variables for secure configuration. A system prompt is defined to ensure GPT‑4.5 understands the tone and style expected. The conversation is built as a list of messages (system + user), which is then sent using the Azure OpenAI client. Finally, we print out the response — this is the core of our Foundry integration, which helps manage model settings and authentication. Sample Code in Python Here’s a Python example using the OpenAI library (configured for Azure OpenAI) to invoke GPT‑4.5 with Foundry integration. import osimport openai# Load environment variables (ensure these are set securely)azure_endpoint = os.getenv("AZURE_OPENAI_ENDPOINT")azure_api_key = os.getenv("AZURE_OPENAI_API_KEY")deployment_name = os.getenv("AZURE_OPENAI_DEPLOYMENT_NAME")# Configure the OpenAI client to use Azure OpenAIopenai.api_base = azure_endpointopenai.api_key = azure_api_keyopenai.api_version = "2024-10-21" # Adjust API version as needed# Define a system prompt for contextsystem_message = ( "You are a friendly and insightful assistant. Please provide detailed and engaging responses, " "using emojis and human-like language when appropriate. 😊")# Prepare the conversation messagesmessages = [ {"role": "system", "content": system_message}, {"role": "user", "content": "Show me an example of invoking GPT-4.5 via Azure OpenAI with Foundry integration."}]# Create a chat completion requestresponse = openai.ChatCompletion.create( model=deployment_name, # This corresponds to the GPT-4.5 deployment in your Azure resource messages=messages, max_tokens=500, temperature=0.7)# Print the generated responseprint("Response from GPT-4.5:")print(response.choices[0].message.content) Explanation: Environment variables are loaded using os.getenv() for secure configuration. The openai module is configured to point to your Azure endpoint and use your API key. We construct a conversation with both a system prompt and a user prompt. The ChatCompletion.create() method sends our […]
↧
