OpenAI
| Model | Best For |
|---|---|
| GPT-5.2 | General coding, fast responses |
| GPT-5 Mini | Quick tasks, cost-effective |
| o4-mini | Reasoning tasks |
| o3 | Advanced reasoning |
| o3-pro | Complex analysis |
| GPT-5.3 | Latest generation, advanced coding |
| GPT-5.2 Pro | Premium coding |
| GPT-5.1 Codex | Code generation |
Anthropic
| Model | Best For |
|---|---|
| Claude Sonnet 4.5 | Balanced coding and analysis |
| Claude Opus 4.5 | Complex reasoning, large contexts |
| Claude Opus 4.6 | Latest generation, advanced reasoning |
| Model | Best For |
|---|---|
| Gemini 2.5 Flash | Fast responses, cost-effective |
| Gemini 3 Pro | Premium analysis |
| Gemini 2.5 Pro | Balanced performance |
DeepSeek
| Model | Best For |
|---|---|
| R1 | Deep reasoning tasks |
Local Models
Run models entirely on your Mac with zero network requests using Ollama or LM Studio. Your code never leaves your machine.Local Models Setup
Learn how to configure Ollama and LM Studio for fully offline AI coding.
OpenRouter
Access 100+ models through a single API key. OpenRouter provides smart routing, cost optimization, and automatic fallbacks between providers.OpenRouter Setup
Configure OpenRouter for access to hundreds of models with one API key.
Recommendations
Quick edits and chat
Quick edits and chat
GPT-5 Mini or Gemini 2.5 Flash — fast response times and low cost. Ideal for simple questions, small edits, and exploratory conversations.
Complex refactoring
Complex refactoring
Claude Sonnet 4.5, GPT-5.3, or GPT-5.2 — strong at understanding large codebases and generating multi-file changes with high accuracy.
Deep analysis
Deep analysis
Claude Opus 4.6, Claude Opus 4.5, or o3-pro — best for architectural decisions, debugging complex issues, and reasoning through intricate logic.
Maximum privacy
Maximum privacy
Local models via Ollama — all inference happens on your Mac. Zero network requests. Ideal for proprietary codebases and regulated environments.