> ## Documentation Index
> Fetch the complete documentation index at: https://docs.aurinfer.com/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Coding Engines

> Comprehensive guide to supported AI coding models, context windows, reasoning capabilities, and multi-engine architecture.

# AI Coding Engines

AurInfer provides access to state-of-the-art AI coding models benchmarked for agentic tool use, precise syntax generation, and multi-file code reasoning.

***

## Supported Models Comparison

| Model                                 | Provider | Context Window | Key Strengths & Badges                                                                               |
| ------------------------------------- | -------- | -------------- | ---------------------------------------------------------------------------------------------------- |
| **GLM 5.2**                           | Z-AI     | 128k tokens    | ⚡ *Fast & Accurate* — Ultra-low latency, ideal for interactive planning and everyday feature coding. |
| **Nemotron 3.5 Lightning 30B**        | NVIDIA   | 128k tokens    | 🧠 *Reasoning* — State-of-the-art agentic tool calling and deep code analysis.                       |
| **Llama 3.3 Nemotron Super 49B v1.5** | NVIDIA   | 128k tokens    | 👑 *Flagship Reasoning* — High reasoning depth for complex multi-file architectural refactors.       |
| **Llama 3.3 70B Instruct**            | Meta     | 128k tokens    | 🌟 *Top Performer* — Excellent general-purpose coding and comprehensive documentation generation.    |
| **Llama 3.1 70B Instruct**            | Meta     | 128k tokens    | 🛡️ *Reliable Polyglot* — Solid multi-language code generation and bug hunting.                      |
| **MiniMax M3 / 01**                   | MiniMax  | 1M+ tokens     | 📚 *Massive Context* — Reads and analyzes extensive codebases in a single turn.                      |
| **GPT-OSS 120B**                      | OpenAI   | 128k tokens    | 🎯 *Code Specialist* — Tuned for clean code implementation, testing, and debugging.                  |
| **Muse Glimmer 30B**                  | Meta     | 32k tokens     | ⚡ *Lightweight & Fast* — Quick helper model for fast summaries and syntax checks.                    |

***

## Multi-Engine Architecture

AurInfer divides AI tasks into specialized roles to ensure maximum speed and reliability:

```mermaid theme={null}
flowchart TD
    User["User Prompt"] --> Studio["Coding Studio"]
    Studio --> Primary["Primary Coding Engine (e.g. GLM 5.2 / Nemotron)"]
    Primary -->|If Rate Limited| Fallback["Fallback Coding Engine (e.g. GPT-OSS 120B)"]
    Studio --> Chat["Chat Engine (e.g. GPT-OSS 120B / Llama 3.3)"]
```

1. **Primary Coding Engine**: The core model responsible for tool calling, reading files, generating code diffs, and debugging.
2. **Fallback Coding Engine**: Automatically takes over if your primary model experiences temporary upstream provider outages or rate limits.
3. **Chat Engine**: Powers fast conversational dialogue during interactive planning.

***

## Switching Engines in Real Time

You can customize your models at any time:

* **Global Character Defaults**: In **GitHub → Settings**, choose your character's default models.
* **Per-Session Overrides**: In the **Coding Studio** or **Active Workspace**, click the **Engine Switcher** dropdown to switch models on the fly. Session overrides only apply to that specific session.