✦ AI

Supported Models

Choose the right AI for your project

Archon supports multiple AI providers. You can switch models at any time from the chat interface.

Cloud models

Provider          Model               Speed    Quality   Cost
─────────────────────────────────────────────────────────────
OpenAI            GPT-4o              ★★★★     ★★★★     $$$
Anthropic         Claude 3.5 Sonnet   ★★★      ★★★★★    $$$
Google            Gemini 1.5 Pro      ★★★★     ★★★★     $$
OpenRouter        Various             varies   varies   varies

OpenAI — GPT-4o

→ Fastest generation speed
→ Great at SwiftUI and modern iOS patterns
→ Best for quick prototypes and simple apps

Anthropic — Claude 3.5 Sonnet

→ Highest code quality and attention to detail
→ Strong at following complex, multi-step instructions
→ Best for production-ready apps with clean architecture

Google — Gemini 1.5 Pro

→ Largest context window (1M tokens)
→ Great for complex, multi-file projects
→ Most affordable cloud option

Local models

Ollama

Run open-source models on your own hardware. No credits needed, no API limits, fully private. Archon connects to your Ollama instance over SSH.

Recommended Ollama models for Archon:

  codellama:13b      — Good at code generation
  llama3:8b          — Fast, general purpose
  deepseek-coder:16b — Strong at Swift/Kotlin
  gemma3:12b         — Balanced speed/quality

Choosing a model

→ Quick prototype — GPT-4o (fastest)
→ Production app — Claude 3.5 Sonnet (best quality)
→ Complex project — Gemini 1.5 Pro (largest context)
→ Offline / privacy — Ollama (local, no credits)
→ On a budget — Ollama or Gemini (cheapest)