OpenCode Zen lists GPT 5.6 Terra but cannot run it

[muenker80]

Description

opencode/gpt-5.6-terra is listed as an available OpenCode Zen model but fails reproducibly at runtime.

Three consecutive minimal requests failed before any generation started. All returned the same Zen gateway error:

Error: No provider available
HTTP 401
{"type":"error","error":{"type":"ModelError","message":"No provider available"}}

The request reached the documented Responses endpoint:

https://opencode.ai/zen/v1/responses
This is not a local missing-credential issue:

• opencode auth list shows OpenCode Zen api as configured.
• opencode models opencode lists opencode/gpt-5.6-terra.
• Zen documentation lists GPT 5.6 Terra and /zen/v1/responses as its endpoint.

Reproduction
opencode run --model opencode/gpt-5.6-terra "Reply exactly: OK"

Expected: OK

Actual: immediate HTTP 401 with ModelError: No provider available; zero input/output tokens.

Plugins

No response

OpenCode version

No response

Steps to reproduce

No response

Screenshot and/or share link

No response

Operating System

No response

Terminal

No response

[github-actions]

This issue might be a duplicate of existing issues. Please check:

• [BUG]: OpenCode Zen free models hy3-free and nemotron-3-ultra-free failing while deepseek-v4-flash-free works #38028: OpenCode Zen free models failing with similar provider errors
• v2: Opencode Zen models return CreditsError while works with v1 #37161: Zen models returning errors in v2 while working in v1
• All paid OpenCode Zen models fail with 'Upstream request failed' — free models work #36506: All paid OpenCode Zen models failing with upstream errors
• opencode-go (Console Go) provider returns 400/401/500 for subscribed models #37056: opencode-go returning 400/401/500 for subscribed models

If your issue is distinct from those (e.g., specific to the gpt-5.6-terra model), please add a comment clarifying what distinguishes it.

[muenker80]

This is distinct from the linked issues:

• It affects the specific paid Zen model opencode/gpt-5.6-terra, not free models ([BUG]: OpenCode Zen free models hy3-free and nemotron-3-ultra-free failing while deepseek-v4-flash-free works #38028).
• This is OpenCode v1.18.0, not the v2 CreditsError regression in v2: Opencode Zen models return CreditsError while works with v1 #37161.
• This is Zen, not OpenCode Go (opencode-go (Console Go) provider returns 400/401/500 for subscribed models #37056).
• The returned error is consistently HTTP 401 ModelError: No provider available, not Upstream request failed as in All paid OpenCode Zen models fail with 'Upstream request failed' — free models work #36506.
• The request reaches the documented GPT 5.6 Terra Responses endpoint and fails before any tokens are generated.

I verified that the local Zen credential is configured and that opencode/gpt-5.6-terra is listed by opencode models opencode.

The three failed requests occurred at 2026-07-23 15:53:34, 15:54:08, and 15:54:15 UTC with Cloudflare Ray IDs a1fbdbd28ff49de8-CMH, a1fbdca7ab0770f3-CMH, and a1fbdcd70fb98f93-BOS.