opencode is an open-source, terminal-based AI coding agent. It supports OpenAI-compatible endpoints and can be pointed directly at the L&S AI Inference API.
More information: API Access
| Model ID | Best for | Context |
|---|---|---|
qwen3.8-27b |
Code generation, editing, agentic coding | 256k |
gpt-oss-120b |
Complex reasoning, planning, synthesis | 128k |
gemma-4-26b-a4b-it |
General chat, light tasks, subagent use | 256k |
All three models support tool/function calling.
Store your key in an environment variable and add it to your shell profile (~/.bashrc, ~/.zshrc, etc.):
export LSIT_API_KEY="your-lsit-api-key"
Then define the L&S endpoint as a custom provider in ~/.config/opencode/opencode.json (global) or opencode.json in your project root:
{
"$schema": "https://opencode.ai/config.json",
"provider": {
"ai": {
"npm": "@ai-sdk/openai-compatible",
"name": "L&S AI Inference",
"options": {
"baseURL": "https://api.ai.college.ucsb.edu/v1",
"apiKey": "{env:LSIT_API_KEY}"
},
"models": {
"qwen3.8-27b": { "name": "Qwen3.8 27B" },
"gpt-oss-120b": { "name": "GPT-OSS 120B" },
"gemma-4-26b-a4b-it": { "name": "Gemma 4 26B A4B" }
}
}
},
"model": "ai/qwen3.8-27b"
}
Models are referenced as ai/<model-id>, for example ai/qwen3.8-27b. Launch with opencode and use /models to switch.
opencode.json — it may cache provider config between sessions.gpt-oss-120b and keep file-level edits on qwen3.8-27b.