22 models across three providers. Use --model <flag> when creating sessions.
Claude Models (Anthropic)
| Model | Flag | Capability | Speed | Context | Notes |
|---|
| Opus 4.6 | opus-4-6 | Most capable | Moderate | 1M | Most capable |
| Sonnet 4.6 | sonnet-4-6 | Balanced | Fast | 1M | Default |
| Haiku 4.5 | haiku-4-5 | Moderate | Fastest | - | Fastest Claude |
| Opus 4.5 | opus-4-5 | High | Moderate | - | |
| Sonnet 4.5 | sonnet-4-5 | High | Fast | - | |
| Opus 4.1 | opus-4-1 | High (Agentic) | Moderate | - | Agentic tasks |
| Opus 4 | opus-4 | High | Moderate | - | |
| Sonnet 4 | sonnet-4 | Moderate | Fast | - | |
Codex/OpenAI Models
| Model | Flag | Capability | Speed | Notes |
|---|
| GPT-5.3-Codex | gpt-5-3-codex | High (Code) | Moderate | Default - coding |
| GPT-5.4 | gpt-5-4 | Flagship | Moderate | |
| GPT-5.4 Mini | gpt-5-4-mini | High | Fast | |
| GPT-5.4 Nano | gpt-5-4-nano | Moderate | Fastest | Cost-effective |
| GPT-5.4 Pro | gpt-5-4-pro | Highest | Slow | Premium |
| o3 | o3 | High (Reasoning) | Moderate | |
| o3-pro | o3-pro | Highest (Reasoning) | Slow | Premium |
| o4-mini | o4-mini | High (Reasoning) | Fast | |
| o3-mini | o3-mini | Moderate (Reasoning) | Fast | |
Gemini Models (Google)
| Model | Flag | Capability | Speed | Context |
|---|
| Gemini 3.1 Pro | gemini-3-1-pro | Highest | Moderate | 1M |
| Gemini 3 Flash | gemini-3-flash | High | Fastest | - |
| Gemini 3.1 Flash-Lite | gemini-3-1-flash-lite | Moderate | Fastest | - |
| Gemini 2.5 Pro | gemini-2-5-pro | High | Moderate | 1M |
| Gemini 2.5 Flash | gemini-2-5-flash | Moderate | Fast | - |
Usage
clauderon create --model opus-4-6 --repo ~/project --prompt "Refactor auth module"
curl -X POST http://localhost:3030/api/sessions \
-H "Content-Type: application/json" \
-d '{"name": "my-session", "repository": "/path/to/repo", "backend": "docker", "agent": "claude-code", "model": "opus-4-6"}'
TUI and Web UI: select model from dropdown during session creation.
Choosing a Model
| Use Case | Recommended Model |
|---|
| Most tasks | Sonnet 4.6 (default) |
| Complex refactoring | Opus 4.6 |
| Quick iterations | Haiku 4.5 |
| Large codebases | Gemini 3.1 Pro (1M) |
| Code generation | GPT-5.3-Codex |
| Deep reasoning | o3 or Opus 4.1 |
| Cost optimization | Haiku 4.5 / GPT-5.4 Nano / Gemini 3.1 Flash-Lite |
API Keys
export ANTHROPIC_API_KEY="sk-ant-..." # Claude
export OPENAI_API_KEY="sk-..." # OpenAI/Codex
export GOOGLE_API_KEY="AIza..." # Gemini
Models cannot be changed mid-session. Recreate the session to switch models.
