Configure OpenAI GPT models in BraceKit.
OpenAI
OpenAI provides GPT models with configurable reasoning levels. The GPT-5.6 family (Sol, Terra, Luna) offers the latest capabilities with improved agentic performance and tool use.
Setup
1. Get an API Key
- Go to platform.openai.com/api-keys
- Sign in or create an account
- Click “Create new secret key”
- Copy the key (starts with
sk-)
2. Configure in BraceKit
- Open Settings
- Select AI Provider tab
- Select OpenAI from the provider dropdown
- Paste your API key in the API Key field
- Select a model from the dropdown
Settings are saved automatically as you type.
Settings are saved automatically when you make changes.
Available Models
GPT-5.6 Family (Recommended)
The latest generation. The gpt-5.6 alias routes to gpt-5.6-sol.
| Model | Best For | Context | Notes |
|---|---|---|---|
| gpt-5.6-sol | Complex professional work | 1.05M | Frontier model, all reasoning levels |
| gpt-5.6-terra | Balanced intelligence & cost | 1.05M | Strong middle tier |
| gpt-5.6-luna | Cost-sensitive, high volume | 1.05M | Fastest, cheapest |
Previous Generation
| Model | Best For | Context | Notes |
|---|---|---|---|
| gpt-5.5 | General purpose | 400K | Previous flagship |
| gpt-5.4 | General purpose, coding | 1.05M | Prior generation, computer use |
| gpt-oss | Open-weight models | 400K | Openly available weights |
Reasoning (Effort Levels)
GPT-5.x models support configurable reasoning effort: minimal, low, medium, and high. The gpt-5.6 alias routes to gpt-5.6-sol.
| Model | Reasoning Levels | Best For |
|---|---|---|
| gpt-5.6-sol | minimal → high | Complex professional work |
| gpt-5.6-terra | minimal → high | Balanced intelligence & cost |
| gpt-5.6-luna | minimal → high | Cost-sensitive, high volume |
Note: Model availability depends on your OpenAI account tier and API access. Some models may require higher tier access.
Features
Reasoning (Effort Levels)
The GPT-5.6 family shows its reasoning process:
┌─────────────────────────────────────┐
│ 🧠 Thinking... ▾ │
├─────────────────────────────────────┤
│ Let me work through this step by │
│ step... │
└─────────────────────────────────────┘
Based on my analysis...This happens automatically when a reasoning level above none is set. The thinking process helps with:
- Complex math and logic problems
- Multi-step coding tasks
- Detailed analysis and planning
Vision (Image Input)
All GPT-5.6, GPT-5.5 and GPT-5.4 models support image analysis:
- Attach an image to your message
- Ask a question about it
- The model analyzes and responds
Supported formats: PNG, JPEG, GIF, WebP
Function Calling
OpenAI models support tool calling for:
- MCP server tools
- Built-in tools (Google Search)
Streaming
All OpenAI models support streaming responses for real-time output.
Model Parameters
Configure in Settings → AI Provider under the Advanced section:
| Parameter | Range | Effect |
|---|---|---|
| Temperature | 0-2 | Higher = more creative, lower = more focused |
| Max Tokens | 1+ | Maximum response length |
| Top P | 0-1 | Controls diversity of word choices |
Note: OpenAI does not support Top K or Thinking Budget parameters. Those are available for Anthropic, Gemini, and Ollama providers.
Recommended Settings
| Use Case | Temperature | Top P | Max Tokens |
|---|---|---|---|
| Code generation | 0.3 | 0.9 | 4096 |
| General chat | 0.7 | 1.0 | Default |
| Creative writing | 1.0 | 1.0 | 8192 |
| Factual Q&A | 0.0 | 0.9 | 2048 |
Pricing
OpenAI charges per token. Check OpenAI pricing for current rates.
Troubleshooting
“Insufficient quota”
- Add credits to your OpenAI account
- Check usage limits at platform.openai.com
“Model not found”
- Verify the model name is correct
- Some models require organization verification
Slow responses
- Higher reasoning levels (xhigh, max) are slower by design — they “think” before responding
- Flagship models (gpt-5.6-sol) take longer for deep reasoning
- For faster responses, use gpt-5.6-luna or gpt-5.6-terra with a low reasoning level
Related
- Anthropic — Claude models
- Gemini — Google models
- Configuration — All settings