Muse Spark 1.2 Contributor works on Command Code but blocked by RegionError on OpenCode Go (same country, same IP) #43692

Summary

muse-spark-1.2-contributor is listed in OpenCode Go's model list and docs (marked "available in select regions"), but calling it from South Korea returns RegionError: This model is not available in your country.

However, the same model works fine on Command Code (another Go-style subscription provider) from the same country and same IP. This inconsistency suggests the region gate may be misconfigured on the OpenCode side.

Reproduction

OpenCode Go (fails)

curl -s -X POST "https://opencode.ai/zen/go/v1/chat/completions" \
  -H "Authorization: Bearer $OPENCODE_GO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"muse-spark-1.2-contributor","messages":[{"role":"user","content":"Reply with exactly: OK"}],"max_tokens":20}'

Response:

{"type":"error","error":{"type":"RegionError","message":"This model is not available in your country."}}

Command Code (works) — same country (South Korea), same IP

curl -s -X POST "https://api.commandcode.ai/provider/v1/chat/completions" \
  -H "Authorization: Bearer $COMMANDCODE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"meta/muse-spark-1.2-contributor","messages":[{"role":"user","content":"Reply with exactly: OK"}],"max_tokens":20}'

Response: HTTP 200, normal chat completion (model: meta/muse-spark-1.2-contributor)

Environment

• Location: South Korea (Seoul, residential IP)
• OpenCode Go model list: muse-spark-1.2-contributor present (via GET /zen/go/v1/models)
• Command Code model list: meta/muse-spark-1.2-contributor present (via GET /provider/v1/models)
• Both subscriptions active (paid)

Expected behavior

If the model is available to Command Code subscribers in South Korea, it should also be available to OpenCode Go subscribers in the same region — or the docs should clearly state which regions are excluded and why the exclusion differs between the two providers that offer the same Meta model.

Additional context

• Related issues: Streaming responses for muse-* models on the Zen gateway never send finish_reason — strict OpenAI-compatible clients enter a retry loop #43379 (streaming finish_reason), Zen gateway does not forward reasoning/thinking stream for Muse Spark 1.2 #43584 (reasoning stream), Muse model fails with “Upstream request failed: Endpoint is unavailable” #43477 (Endpoint unavailable), [Opencode Go] Frequent jitter / possible truncation when calling opencode-go/muse-spark-1.2 via Codex (opencodex proxy, openai-chat) #43680 (jitter/truncation)
• The RegionError appears before any model invocation — it is a gateway-level region check, not a model error.