Standard
For everyday coding Tasks, bug fixes, and small features.
Models
Choose how your Tasks run.
Alloyqa uses a Codex-based coding harness to work on your code, run checks, and prepare pull requests in isolated Cloud environments.
Task modes
Deep
For complex Tasks that need a more capable model or additional reasoning.
Available models
Models currently available in Alloyqa.
Alloyqa Default
Use Alloyqa without connecting an API key.
| Mode | Model | Reasoning |
|---|---|---|
| Standard | Gemini 3.8 Flash | Medium |
| Deep | Gemini 4 Argon | Medium |
OpenAI
Use your own OpenAI API key.
| Mode | Model | Reasoning |
|---|---|---|
| Standard | GPT-5.6 Terra | Medium |
| Deep | GPT-6.1 Sol | High |
Bring your own API key
Use your own OpenAI API key.
Connect your OpenAI API key in Settings and choose OpenAI when creating a Task.
Your key is encrypted and kept separate from the coding environment.
Billing
Model usage and Cloud execution.
Alloyqa Default
Your Alloyqa Cloud credits cover both model usage and Cloud execution.
OpenAI
OpenAI bills you directly for model usage. Alloyqa charges $0.01 per started minute of Cloud execution, up to $1.20 per Task.
Cloud execution
Each Task runs independently.
Each Task runs independently in the Cloud, on its own branch.
Alloyqa makes changes, runs available checks, and opens a pull request for review.
You can review the changes and leave feedback when you're ready.