Skip to content

OpenCode Go models return "Insufficient account funds" with active subscription and 0% usage #51424

Closed as not planned

Description

@karindoushoten-wq

Description

I subscribed to OpenCode Go today.

The Go subscription is active and the dashboard shows:

  • Rolling usage: 0%
  • Weekly usage: 0%
  • Monthly usage: 0%
  • Extra Usage: OFF
  • Available credit: $0.00

I logged into the CLI successfully with opencode console login.

I then selected Kimi K2.7 Code from the model list and sent a simple request, but it repeatedly returns:

Upstream request failed: Insufficient account funds

The Go usage remains at 0%.

Plugins

None

OpenCode version

1.18.32

Steps to reproduce

  1. Subscribe to OpenCode Go.
  2. Run opencode console login and complete authentication successfully.
  3. Start OpenCode.
  4. Open /models.
  5. Select Kimi K2.7 Code under Personal / OpenCode.
  6. Send a simple prompt such as Reply with only: OK.
  7. The request fails with Upstream request failed: Insufficient account funds.
  8. Check the Go dashboard. Usage remains at 0%.

Screenshot and/or share link

Image

Operating System

No response

Terminal

No response

Activity

  1. github-actions commented on Sep 26, 2026

    @github-actions
    Contributor

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

    • #50087: OpenCode Go subscription is active, but no Go models appear in /model.
    • #50611: Console sign-in is associated with Go model access failures and Insufficient account funds.
  2. kennyzcx commented on Sep 28, 2026

    @kennyzcx

    Getting the same issue, please let me know if you find a solution.

  3. stunglebungus commented on Sep 28, 2026

    @stunglebungus

    Confirming this is still broken — additional data points that might help narrow it down:

    1. Tried the desktop: "Connect OpenCode Go" runs Console sign-in instead of the Go API-key flow, then reports success and fails with "Insufficient account funds" #50611 workaround (registering the key directly via the local API: POST /api/integration/opencode-go/connect/key) — it worked once, creating a proper opencode-go credential distinct from the generic opencode/Console one. Model calls still failed with Insufficient account funds.
    2. The next day, generated a brand new API key and repeated the exact same workaround. This time the registration call itself failed server-side with:
      {"name":"UnknownError","data":{"message":"Unexpected server error. Check server logs for details.","ref":"err_2c57926c"}}
      Local server log: error="Key method not found: opencode-go"
    3. Separately, a third-party tool (Orca) that polls the OpenCode Go usage API independently reports a third, different error for the same account: "The OpenCode Go API key was rejected", failureKind: "stale-token".

    Three different failure modes (insufficient funds → key-method-not-found → stale-token) across two days on the same account with no client-side changes points at something inconsistent on the backend entitlement/credential linkage rather than a per-user funding issue. Subscription active, 0% usage throughout.

  4. stephenmcconnachie commented on Sep 29, 2026

    @stephenmcconnachie

    I am getting the same issue when using opencode 2 TUI. Using the console auth or the API auth both have the same outcome - 'insufficient account funds' error when using any paid model, when Go levels are not at maximum for day / week / month.

    This is only happening for new logins to OpenCode and using the console or API method to add OpenCode Go. Where I am already authenticated against the OpenCode Go service with API key, my prompts are working without errors.

    Successful call, uses the Product: go and unsuccessful cal luses the Product: standard - screegrabs below. So new logins are associating with standard, and old logins were associated with go.

    Image Image
  5. absurdother commented on Sep 29, 2026

    @absurdother

    Same error. No fix?! Please.

  6. stephenmcconnachie commented on Sep 30, 2026

    @stephenmcconnachie

    Now getting a different error shape and behaviour - sometimes the OpenCode Go endpoint responds, but sometimes returns 400.

    Please look into this, it's making the service unusable for any real work.

    Image
  7. rubslopes commented on Oct 2, 2026

    @rubslopes
  8. rbmr commented on Oct 8, 2026

    @rbmr

    Update: this was not a bug on the Go side for me. I was using the plain opencode/... provider, which bills the credit balance ($0.00), so "Insufficient account funds" was the correct result. opencode-go/... uses the Go subscription and works.

    If you see this error with an active Go plan, check that:

    • You picked a model listed under OpenCode Go in /models, with an ID like opencode-go/deepseek-v4.1-flash. opencode/deepseek-v4.1-flash is a different provider and bills credits.
    • Your key is connected with /connect → OpenCode Go.
    • Your default model in opencode.json doesn't still point at an opencode/... model.

    Raw API calls need the Go endpoint https://opencode.ai/zen/go/v1/chat/completions and an x-opencode-session header. I tested it with opencode v2.0.22. The /inference/openai/v1 path from the docs example bills credits too.

    I can't say this is the cause for everyone in this thread. Some reports above describe other symptoms.

  9. GallardoGH commented on Oct 9, 2026

    @GallardoGH

    Same symptom from a different client (OpenClaw 2026.9.8, Windows). Requests to https://opencode.ai/zen/go/v1/chat/completions intermittently return:

    402 {"error":{"type":"server_error","message":"Upstream request failed: Insufficient account funds"}}

    while the same API key + same model + same endpoint succeed seconds later.

    Data (2026-10-07/08, Go subscriber, Go quota nowhere near limits):

    • Go endpoint /zen/go/v1: 301× HTTP 200 vs 7× 402 on 10-07 (all 402s clustered 21:23–21:53 UTC); 64× 200 vs 1× 402 on 10-08 (14:20 UTC). ~98% success between bursts.
    • Zen endpoint /zen/v1 with the same key returns the byte-identical 402 body on every call (Zen balance is $0.00) — so when Go calls fail, the response is indistinguishable from the "no Zen balance" error, which matches this issue's "billed against Zen balance instead of Go quota" hypothesis.
    • A local diagnostic probe of ours through the Go provider fails 6/6 on deepseek-v4-pro with that exact 402, while direct calls to /zen/go/v1 with deepseek-v4-pro return 200 minutes apart (same key, same model).
    • Free/promotional models (e.g. longcat-2.5-preview-free) return 200 at all times.

    Repro cadence: bursts of ~30 min, then clean for hours. Happy to share precise request timestamps if useful.

  10. thdxr commented on Oct 11, 2026

    @thdxr
    Member

    Thanks for the report! OpenCode 2.0 is now out and development is focused on V2, so we're closing issues filed against OpenCode 1. Please upgrade (opencode upgrade --major, details: https://anoma.ly/notes/opencode-2-0), and if this is still relevant on 2.0, reopen this issue or open a new one with your 2.0 version.

    — from 𝕺𝖕𝖊𝖓𝕮𝖔𝖉𝖊

  11. gambikules commented on Oct 15, 2026

    @gambikules

    I have this problem

  12. Mat-Ice-QC commented on Oct 17, 2026

    @Mat-Ice-QC

    i still have this problem too :(

  13. FallDownTheSystem commented on Oct 17, 2026

    @FallDownTheSystem

    EDIT: Make sure you have the "OpenCode Go" provider selected when choosing models, not the OpenCode/Personal one.

  14. sp80808 commented on Oct 21, 2026

    @sp80808

    still having this issue

  15. KappelM commented on Oct 27, 2026

    @KappelM

    I have this issue! wtf why is it not fixed yet?

  16. w3bKodr commented on Nov 2, 2026

    @w3bKodr

    I'm also having this issue, any resolutions?

  17. pianopia commented on Nov 2, 2026

    @pianopia

    In my case this was related to the API key, not the Go subscription itself.

    Two things I hit:

    1. An expired API key — regenerating/updating the key fixed the Insufficient account funds error.
    2. The key also seemed to disappear earlier than expected. The default expiry looked like 90 days, but mine was already gone sooner than that.

    So if usage is still at 0% with an active Go sub, it's worth checking whether the API key was rotated, expired, or somehow removed, before digging into billing.

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

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions