Run OpenCode on OAIRoute
OpenCode reads providers from opencode.json. Register OAIRoute as an OpenAI-compatible provider and every model it lists shows up in the model switcher.
Configuration
{
"provider": {
"oairoute": {
"npm": "@ai-sdk/openai-compatible",
"options": {
"baseURL": "https://api.oairoute.com/v1",
"apiKey": "YOUR_OAIROUTE_KEY"
},
"models": {
"claude-opus-5": { "name": "Claude Opus 5" },
"gpt-5.6": { "name": "GPT-5.6" }
}
}
}
}opencode and pick the provider with the model switcher.Why route it through OAIRoute
Terminal-native, no GUI to maintain
The whole setup is a JSON block. Copy it, swap the key, switch models with one keystroke.
Any vendor from the same picker
Claude, GPT, Gemini and DeepSeek are all one model string apart in the switcher.
Works with the AI SDK provider
Underneath it uses @ai-sdk/openai-compatible, so streaming and tool calls behave as expected.
Setup
- 1
Create a key
Register at api.oairoute.com and create a key on the token page.
- 2
Write the config
Save the JSON below to ~/.config/opencode/opencode.json.
- 3
Verify
Run `opencode` and pick the provider with the model switcher.
Models worth trying
| Model ID | Name | Use it for | USD per 1M in / out |
|---|---|---|---|
claude-opus-5-5 | Claude Opus 5.5 | Default for hard terminal tasks. | $0.20 / $1.00 |
gpt-6.1-sol | GPT-6.1 Sol | Good general-purpose default. | $0.10 / $0.50 |
kimi-k3 | Kimi K3 | Strong on Chinese-language tasks. | $0.90 / $4.50 |
Prices are per million tokens, input / output, at the rates published in the model catalogue.
Questions
Where does OpenCode look for config?
~/.config/opencode/opencode.json on macOS and Linux. On Windows it follows your XDG config directory.
Do I need to list every model?
No. You can add just the ones you plan to use; the provider will still pass other model IDs through.
Streaming or tool calls break?
Check that the npm package is @ai-sdk/openai-compatible and that baseURL ends in /v1.