GPT-6.1 Sol for C# Developers: The Model I'd Test on Every Pull Request

Description and use

This guide explains how to integrate GPT‑6.1 Sol into a .NET project to automatically analyze pull requests, using Microsoft.Extensions.AI and Azure.Identity.

Problem

Integrate GPT‑6.1 Sol into a .NET CI pipeline to automatically analyze and suggest changes for every pull request, without adding Python tooling.

Outcome

The workflow enables a .NET team to run GPT‑6.1 Sol on every pull request, automatically generating suggestions and measuring token usage, all within the existing C# ecosystem.

Workflow

• 1. Deploy "gpt-6.1-sol" from the Foundry catalog and note endpoint and deployment name.
• 2. Assign the Cognitive Services OpenAI User role to the identity that will run the code.
• 3. Configure local development with Azure.Identity DefaultAzureCredential (az login or VS Code sign‑in).
• 4. Create a .NET console project (e.g., dotnet new console -n Gpt-6.1-Sol-CSharp-Demo).
• 5. Add required NuGet packages: Azure.AI.OpenAI 2.1.0, Azure.Identity 1.21.0, Microsoft.Extensions.AI 10.10.0, Microsoft.Extensions.AI.OpenAI 10.10.1, Microsoft.Extensions.Configuration.UserSecrets 10.0.12.
• 6. Initialize user secrets and store AZURE_AI_ENDPOINT and AZURE_OPENAI_DEPLOYMENT values.
• 7. Write C# code that builds a ConfigurationBuilder, retrieves secrets, creates AzureOpenAIClientOptions with a 3‑minute timeout, and constructs an IChatClient.
• 8. Use the client to send prompts to GPT‑6.1 Sol, capture token usage, and output results to the console or CI logs.
• 9. Integrate the console app into the CI pipeline (e.g., as a GitHub Action or Azure DevOps task) to run on every PR.
• 10. Compare token usage and output between GPT‑6 Sol and GPT‑6.1 Sol before switching traffic.
• 11. Deploy the chosen model to production, ensuring role assignments are in place.
• 12. Monitor cost and latency, adjusting prompts or token limits as needed.

Evidence

Deploy "gpt-6.1-sol" from the Foundry catalog and note endpoint and deployment name.

Source

Assign the Cognitive Services OpenAI User role on the resource to whoever runs the code: your own account for local development, a managed identity in production.

Source

Create the project. I compiled every sample in this post against the package versions below on .NET 10; pinning them keeps the code from drifting as Microsoft.Extensions.AI evolves.

Source

The endpoint and deployment name live in .NET user secrets, which are stored in your user profile outside the project folder, so they never end up in source control.

Source