Skip to content

All OpenCode Zen free models blocked for three days — HTTP 429 on Windows #58900

Closed as not planned

Description

@Gunkov

Hello OpenCode Support,
For the past three days, all free models available to me through OpenCode Zen have been blocked in the Windows desktop app. Switching between models does not help. Every attempt returns “Free usage exceeded, subscribe to Go,” followed by a long retry countdown.
This affects DeepSeek V4.1 Flash Free, MiMo V2.6 Flash Free, GLM-5.3 Flash Free, Beam Free, and other free Zen models.
The issue persists after upgrading OpenCode Desktop from version 1.18.31 to 1.19.4 and restarting the application. A previously configured local model gateway was removed, and OpenCode now connects directly to Zen. Windows proxy settings are disabled.
A diagnostic request using the official Desktop 1.19.4 bundled backend with an empty project returned:

  • Time: October 25, 2026, 05:46:59 UTC
  • Model tested: deepseek-v4.1-flash-free
  • HTTP status: 429
  • Error type: FreeUsageLimitError
  • Message: “Free usage limit reached.”
  • Retry-After: 65581 seconds, pointing to October 26, 2026, at 00:00 UTC
    The local message history shows no successful Zen responses on October 24 or October 25 before this diagnostic test.
    Could you please investigate why all free Zen models remain blocked? Is this an expected usage restriction, a shared-IP limit, or a quota-accounting issue? Please clarify which limit is being applied and how access can be restored.
    I have attached a diagnostic report without API keys or passwords. Please let me know if you need additional details privately.
    Thank you.

opencode-zen-diagnostic-2026-10-25.md

Activity

  1. opencode-agent commented on Oct 25, 2026

    @opencode-agent
    Contributor

    Thanks for the detailed report, @Gunkov. Being locked out of every free Zen model for three days, even after removing the gateway and connecting directly, is frustrating.

    The FreeUsageLimitError (HTTP 429, with Retry-After pointing to 00:00 UTC) is the Zen free-usage limit, which is enforced on the server for your account or connection. Only the team can check which limit is being applied and whether your usage was counted correctly. Billing and account support isn't handled in GitHub issues, so please email help@anoma.ly with the details: the timestamps, the model you tested and the diagnostic report you attached. If you've already emailed, the team will get to it, especially now that we've seen it here.

    Please keep email addresses, account identifiers, API keys and any payment details out of this public issue. Put them only in the email.

    These open issues report similar free-tier 429s and may be useful to follow:

    The bot will close this issue in 1 day, since it needs to go through the support email.

  2. luciano350hp commented on Oct 25, 2026

    @luciano350hp

    Same problem here. Windows, OpenCode Desktop v1.18.33.

    • Every free model (Big Pickle, DeepSeek Flash Free, Beam Free, Flint Free) fails with
      "Free usage exceeded, subscribe to Go" plus a long retry countdown.
    • Last working use was around Oct 22. Since then, a new session and a different model
      make no difference.
    • Screenshot on Oct 24 at 13:17 (UTC-3) showed "retrying in 27750s (attempt 1)".
      That is about 7h42m, which lands at 00:00 UTC. That matches the daily reset reported
      in this issue, but the block continued afterwards.
    • The old session was very large (638 messages, ~5.2M input tokens + 29M cached reads
      on Big Pickle). A new small session also fails.
    • I am a student using the free tier for a university project (OpenCode + OpenSpec + MCP).

    Is this an expected limit, a per-IP limit, or a quota accounting problem? I am also
    emailing help@anoma.ly.

  3. opencode-agent commented on Oct 26, 2026

    @opencode-agent
    Contributor

    Closing this now, since Zen free-usage limits and account questions are handled by email rather than in GitHub issues. @Gunkov, if your free models are still blocked or you want to know which limit is being applied, please email help@anoma.ly with your timestamps and diagnostic details. Please keep account identifiers, API keys and payment details out of this public issue. Thanks for the thorough report!

  4. muham65 commented on Oct 26, 2026

    @muham65

    This is the third time my free usage limit has done this. The countdown keeps resetting to tens of thousands of seconds instead of actually resetting, and all free models return "429 Rejected" immediately. I've already waited for the limit to reset multiple times, but the same thing keeps happening. Is anyone else experiencing this? Is there an actual fix?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions