Jon Galloway, Pamela Fox, and Dan Wahlin present a comprehensive guide for developers on using GPT-5 within Microsoft platforms. The post covers integration in GitHub Copilot, Azure AI Foundry, Copilot Studio, VS Code, and .NET SDK, including technical examples and best practices.

GPT-5 Integration for Microsoft Developers: Tools, APIs, and Code Samples

Stay at the forefront of AI development with a detailed rundown of GPT-5 support and integration options across Microsoft’s developer ecosystem, brought to you by Jon Galloway, Pamela Fox, and Dan Wahlin.

Announcement Overview

OpenAI introduced GPT-5, focusing on improved reasoning, context handling, unified capabilities, and multimodal features.
Microsoft rolled out GPT-5 support immediately in tools and services relevant to developers.

Where to Use GPT-5 in the Microsoft Ecosystem

GitHub Copilot

Richer suggestions and enhanced code completion using GPT-5, especially on larger codebases or multifile refactoring.
Integrated into VS Code, Visual Studio, JetBrains IDEs, Xcode, Eclipse.
“GPT-5 mini” available in all Copilot plans.

AI Toolkit in Visual Studio Code

Experiment with GPT-5 via AI Toolkit: connect to GitHub Models or Azure AI Foundry, run playgrounds, and scaffold new integrations.
Supports both cloud and OSS/local backends.

Azure AI Foundry

Enterprise-grade GPT-5 model hosting.
Advanced features: structured output, model routing, agentic, and reasoning tasks.
Initially available in East US 2 and Sweden Central (requires registration for some models).

GitHub Models Marketplace

Browse, experiment, and use the GPT-5 family (full, mini, nano, chat variants), including CLI access for prompt engineering and workflow automation.

Copilot Studio

Makers can select GPT-5 or GPT-5 chat/reasoning models for advanced agent orchestration.

OpenAI .NET SDK

Official .NET SDK supports GPT-5 via Responses API with features like streaming and configurable reasoning efforts.
NuGet package available.

Code Examples with GPT-5

C# Example – Streaming with Reasoning Effort

using OpenAI.Responses;
OpenAIResponseClient client = new(
    model: "gpt-5",
    apiKey: Environment.GetEnvironmentVariable("OPENAI_API_KEY")
);
await foreach (var update in client.CreateResponseStreamingAsync(
    userInputText: "Explain beta-reduction in lambda calculus.",
    new ResponseCreationOptions {
        ReasoningOptions = new ResponseReasoningOptions {
            ReasoningEffortLevel = ResponseReasoningEffortLevel.High,
        },
    })) {
    if (update is StreamingResponseContentPartDeltaUpdate delta) {
        Console.Write(delta.Text);
    }
}

Python Example – Fine-tuned Reasoning and Verbosity

import os
import openai
from azure.identity import DefaultAzureCredential, get_bearer_token_provider
client = openai.AzureOpenAI(
    api_version=os.environ["AZURE_OPENAI_VERSION"],
    azure_endpoint=os.environ["AZURE_OPENAI_ENDPOINT"],
    azure_ad_token_provider=get_bearer_token_provider(DefaultAzureCredential(), "https://cognitiveservices.azure.com/.default"),
)
response = client.chat.completions.create(
    model=os.environ["AZURE_OPENAI_DEPLOYMENT"],
    messages=[{"role": "user", "content": "Explain beta-reduction in lambda calculus."}],
    reasoning_effort="minimal",
    verbosity="low"
)
print(response.choices[0].message.content)

JavaScript Example – Structured Output

import { AzureOpenAI } from "openai";
import dotenv from "dotenv";
dotenv.config();
const endpoint = process.env.AZURE_INFERENCE_ENDPOINT;
const key = process.env.AZURE_INFERENCE_KEY;
const deployment = process.env.AZURE_OPENAI_DEPLOYMENT || "gpt-5";
const client = new AzureOpenAI({
  endpoint,
  apiKey: key,
  apiVersion: "2025-01-01-preview",
  deployment,
});
const schema = {
  name: "math_explanation",
  schema: {
    type: "object",
    properties: {
      steps: { type: "array", items: { type: "string" } },
      answer: { type: "number" },
    },
    required: ["steps", "answer"],
    additionalProperties: false,
  },
  strict: true,
};
const result = await client.chat.completions.create({
  model: deployment,
  messages: [
    { role: "system", content: "Return JSON only." },
    { role: "user", content: "What is 23 * 7? Show your steps." },
  ],
  response_format: { type: "json_schema", json_schema: schema },
});
const content = result.choices[0].message?.content ?? "{}";
const data = JSON.parse(content);
console.log("Steps:", data.steps);
console.log("Answer:", data.answer);

Evaluation and Model Comparison

Use Azure AI Evaluation SDK and Microsoft.Extensions.AI.Evaluation for benchmarking and regression tests.
Prefer structured output (JSON schemas) for automated quality checks.
Explore the RAG chat app with Azure OpenAI and Azure AI Search (Python) for end-to-end deployment and evaluation workflows.

Community and Reference Projects

Pamela Fox: GPT-5: Will it RAG?
Anthony Shaw: GitHub Models CLI, LLM in GitHub Actions
Burke Holland: Vibe coding with GPT-5 in VS Code

Quick Start

Use GPT-5 in VS Code
Use GPT-5 in Visual Studio
Access all API and model docs directly from Microsoft product links above.

Microsoft’s early rollout of GPT-5 unlocks powerful AI tools and deep integration for developers across its ecosystem. Use the above code samples and resources to start building and evaluating GPT-5 applications today.

This post appeared first on “Microsoft DevBlog”. Read the entire article here