[FEATURE]: Add Qwen3.8-27B

Feature hasn't been suggested before.

• I have verified this feature I'm about to request hasn't been suggested before.

Describe the enhancement you want to request

Requesting the inclusion of the Qwen3.8-27B open-weight model in the OpenCode Go subscription catalog (opencode-go/...). Adding this mid-sized model will give users a high-performing, cost-efficient option for everyday coding tasks, local-agent parity, and fast multi-file refactoring within the OpenCode ecosystem.

Proposed Solution
Integrate Qwen3.8-27B into the OpenCode Go managed model catalog.

Expose the endpoint under the explicit runtime reference (e.g., opencode-go/qwen3.8-27b).

Provide standard OpenAI/Anthropic API compatibility for seamless integration with the OpenCode CLI, IDE extensions, and third-party coding agents.

Key Reasons & Benefits

• Optimal Sweet Spot for Speed & Accuracy
  At 27 billion parameters, Qwen3.8-27B strikes an ideal balance between low latency and strong reasoning/code-generation capabilities. It outperforms smaller 7B/14B models on complex coding benchmarks while remaining significantly faster and cheaper to serve than dense 70B+ models.

• Strong Agentic & Multi-File Capabilities
  The Qwen series is natively tuned for tool calling, precise repository context understanding, and file editing. Adding Qwen3.8-27B directly enhances OpenCode's agentic workflows (e.g., automated refactoring, test generation, and terminal command execution).

• Diversity in Open-Source Model Options
  Users who prefer the Qwen architecture for specific languages, long context retention, or regional strengths currently lack a mid-weight Qwen option in OpenCode Go. Adding Qwen3.8-27B fills a key gap between ultra-light models and massive, heavy-tier models.