Console Go (zen/go/v1): deepseek-v4-flash returns RegionError requiring China opt-in, no programmatic opt-in path

Summary

deepseek-v4-flash is listed in GET /zen/go/v1/models but every completion request fails with a RegionError 403 that demands a manual opt-in via a workspace link. API-only consumers (agents/gateways) cannot opt in — the error only contains a browser URL.

Repro

curl -s https://opencode.ai/zen/go/v1/chat/completions \
  -H "Authorization: Bearer <key>" -H "Content-Type: application/json" \
  -d '{"model":"deepseek-v4-flash","messages":[{"role":"user","content":"hi"}],"max_tokens":10}'

Response (reproduced 2026-07-31, multiple times)

{"type":"error","error":{"type":"RegionError","message":"The latest version of this model is only available hosted in China and requires explicit opt in: https://opencode.ai/workspace/wrk_01KJHRCVJTA9W5D147MYK8K5JD/go"}}

Impact

• Model appears healthy in /models list → agents pick it as default, then every turn fails (403, non-retryable)
• No API-surface opt-in exists; the only path is a human clicking a workspace link
• HTTP status observed as 403 with error.type: RegionError (not a standard OpenAI-style error code)

Expected

• Either an API opt-in mechanism (e.g. header/body flag, workspace setting changeable via API), or
• Model excluded from /models for regions without opt-in, or
• Error body documenting how to opt in programmatically

Environment

• Endpoint: https://opencode.ai/zen/go/v1 (Console Go)
• Auth: sk-... Bearer key
• All other listed models work on same key (tested 2026-07-31): kimi-k3, glm-5.2, qwen3.7-plus, deepseek-v4-pro, gpt-5.6-luna, grok-4.5, etc.