Feature: Add built-in support for Qwen3.6 GGUF models

[MaINDotnet-Bot]

⚠️ AI-proposed via MaIN.Docs. Auto-closed in 3 days unless a contributor engages.

Problem Statement

The Qwen3.6 series models are a new and capable family of open models. Currently, users who want to use the GGUF versions with MaIN.NET's Self backend must manually register them using new GenericLocalModel(...). This is functional but not as convenient as using a built-in, "out-of-the-box" model constant.

Proposed Solution

To improve the user experience and formally add support for these models, we should integrate them into the core library.

1. Add Model Constants: Add new constants for the Qwen3.6 series models to the MaIN.Domain/Models/Models/Generics.cs file. For example:

   • Qwen3_6_0_5B_Instruct
   • Qwen3_6_7B_Instruct
   • Qwen3_6_32B_Instruct
   • Qwen3_6_70B_Instruct

2. Add Model Context: Add corresponding entries to MaIN.Domain/Models/ModelContext/ModelContext.cs, defining their context window size (e.g., 32768 tokens) and model type (ModelType.Local).

3. Add a Sample: Create a new sample project in the samples/ directory (e.g., samples/06-Local-Qwen3.6-Chat) that demonstrates how to:

   • Configure the Self backend.
   • Use AIHub.Models.EnsureModelDownloadedAsync() with one of the new Qwen3.6 constants.
   • Run a simple chat completion using the downloaded model.

Acceptance Criteria

• New model constants for the Qwen3.6 series are available under MaIN.Domain.Models.Models.Generics.
• AIHub.Models.GetModelContext() returns the correct metadata for the new models.
• The new sample project compiles and runs successfully, demonstrating a chat interaction with a Qwen3.6 GGUF model via the Self backend.
• The main README.md is updated to list Qwen3.6 in the supported local models section.