You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Go plan hits 100%, blocks 12h — "Use balance" enabled but $39.89 Zen balance never used, on any model #42938
My OpenCode Go subscription reaches 100% monthly usage and then blocks the "Big Pickle" model for 12 hours. I have "Use balance" enabled in the console, and I have $39.89 available on my Zen balance. According to the docs (opencode.ai/docs/go/), Go should fall back to the Zen balance automatically once the Go limit is reached instead of blocking requests.
This has been happening for approximately two weeks. The Zen balance is never touched — not with the default model, and not when I manually switch to a different model (e.g. GLM). The $39.89 remains completely unused while I'm blocked for 12h at a time.
This looks like the same underlying issue reported in #28129 (upstream provider balance depleted, causing "Insufficient balance" even when the user's own subscription/balance is fine) and #32278 (workspace balance status not refreshing correctly).
Steps to reproduce
Use OpenCode Go until monthly usage hits 100%.
Confirm "Use balance" is enabled in the console (billing page).
Confirm Zen balance shows $39.89 available.
Continue using any model, including switching away from Big Pickle to another provider (GLM).
Requests block for ~12h; Zen balance stays at $39.89, untouched.
Expected behavior Once Go's monthly limit is reached, requests should draw from the Zen balance automatically, as described in the Go documentation, regardless of which model is selected.
Actual behavior Requests block for 12h; the $39.89 balance is never deducted, on any model.
Request Please confirm whether this is the same upstream-provider-balance issue as #28129, and unblock the Zen balance fallback so it can be used with a model beyond Big Pickle while my Go quota is exhausted.
Plugins
NONE
OpenCode version
1.18.18
Steps to reproduce
Use OpenCode Go until monthly usage hits 100%.
Confirm "Use balance" is enabled in the console (billing page).
Confirm Zen balance shows $39.89 available.
Continue using any model, including switching away from Big Pickle to another provider (GLM).
Requests block for ~12h; Zen balance stays at $39.89, untouched.
This issue might be a duplicate of existing issues. Please check:
OpenCode 服务故障反馈(go 套餐用尽后,无法转到zen 余额) #39470: Go subscription exhausted, Zen balance fallback not working (HTTP 500 on all models after Go quota used up, "Use balance" enabled but balance not drawn)
Confirming this matches #39470 exactly — same symptom (Go quota hit, "Use balance" enabled, Zen balance never drawn, blocked 12h). See #42938 for a fresh repro with share link if useful.
Thanks for confirming @n-osennij — sounds like the same root cause, just a different trigger point (losing the whole monthly allowance vs. the Zen balance never being drawn at all in my case). Checked #39470 and #33495 and they describe the identical symptom: "Use balance" shows enabled, but the fallback never fires once the Go quota is spent, so you get blocked/HTTP 500 instead of drawing from the paid Zen balance. With three independent reports pointing at the same mechanism, it'd be worth consolidating so it gets the priority the underlying bug deserves. Happy to share logs/screenshots if that helps triage.
Following up — it's been 13 days since my last comment here (confirming this matches #39470), still no response from a maintainer. @fwang, since you're assigned — any update?
Status unchanged: Zen balance still shows $39.89, unused, "Use balance" still enabled, Go quota still blocks all models at 100%. Switching to grok-4.5 as a workaround didn't help either — makes sense, since this is the Zen-balance-fallback bug, not a model-specific one.
Also noting another duplicate that's shown up since: #45324, same "balance available but blocked" symptom. That's at least 3 independent reports (this issue, #39470, #45324) plus n-osennij's comment above, all pointing at the same fallback mechanism never firing. Could this get prioritized, or is there a known workaround beyond waiting out the 12h block?
it's been fkin almost 1 mouth but still no man cares about this, I'm driving crazy using this damn api with openclaw, it never use my remains, but only pops 401 to prevent me the api call, very GOOD JOB!
Two new reports here since my last update, so a short status for anyone landing on this thread.
On 30/08 OpenCode Support told me by email that the block comes from a separate account-level monthly usage limit (in my case $40), and that "Use balance" does not override it. According to them, raising that limit in the workspace Billing settings lets requests fall back to the Zen balance. I haven't tested that myself: I waited for the monthly window to reset (~02/09) and it worked again. It might also explain the 401s mentioned above, although those could be a different problem.
So the fallback exists, but nothing in the UI says the toggle is capped by that limit, and the error doesn't say which limit you hit. That is what still needs fixing here: either "Use balance" should kick in once the Go quota is used, or the UI and the error should name the monthly limit and link to where it can be changed.
@khaosdoctor, thanks for opening #51682. The free-models block looks like the same provider-wide lock; I've added my data there.
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.
Update from OpenCode 2.0.29 (Windows). My original scenario (Go quota exhausted plus "Use balance") can't be tested any more: the workspace no longer has a Go subscription.
What I can confirm: requests through the standard endpoint do draw from the workspace credit. Sixteen Gemini 4 Argon requests on 16 Oct returned 200 and were billed.
What still hurts: requests for Grok 4.5/4.7 and GPT-6 Luna through /inference/go/ returned 403 inference_failed (13–17 Oct), at no cost. The only message is "The inference request failed." I assume the missing Go subscription is the cause, but nothing says so. A clearer error would have saved a lot of debugging.
I won't reopen this. Thanks for the pointer to 2.0.
Activity
github-actions commentedon Aug 16, 2026
This issue might be a duplicate of existing issues. Please check:
If your issue is the same underlying problem, adding a thumbs-up or comment on those issues may help prioritize the fix.
CinematicEnciclopedia commentedon Aug 16, 2026
Confirming this matches #39470 exactly — same symptom (Go quota hit, "Use balance" enabled, Zen balance never drawn, blocked 12h). See #42938 for a fresh repro with share link if useful.
n-osennij commentedon Aug 17, 2026
I agree. Except in my case, I lost the entire monthly limit, even though I had about 30% left in reserve.
CinematicEnciclopedia commentedon Aug 17, 2026
Thanks for confirming @n-osennij — sounds like the same root cause, just a different trigger point (losing the whole monthly allowance vs. the Zen balance never being drawn at all in my case). Checked #39470 and #33495 and they describe the identical symptom: "Use balance" shows enabled, but the fallback never fires once the Go quota is spent, so you get blocked/HTTP 500 instead of drawing from the paid Zen balance. With three independent reports pointing at the same mechanism, it'd be worth consolidating so it gets the priority the underlying bug deserves. Happy to share logs/screenshots if that helps triage.
CinematicEnciclopedia commentedon Aug 30, 2026
Following up — it's been 13 days since my last comment here (confirming this matches #39470), still no response from a maintainer. @fwang, since you're assigned — any update?
Status unchanged: Zen balance still shows $39.89, unused, "Use balance" still enabled, Go quota still blocks all models at 100%. Switching to grok-4.5 as a workaround didn't help either — makes sense, since this is the Zen-balance-fallback bug, not a model-specific one.
Also noting another duplicate that's shown up since: #45324, same "balance available but blocked" symptom. That's at least 3 independent reports (this issue, #39470, #45324) plus n-osennij's comment above, all pointing at the same fallback mechanism never firing. Could this get prioritized, or is there a known workaround beyond waiting out the 12h block?
bitingass commentedon Sep 7, 2026
it's been fkin almost 1 mouth but still no man cares about this, I'm driving crazy using this damn api with openclaw, it never use my remains, but only pops 401 to prevent me the api call, very GOOD JOB!
khaosdoctor commentedon Sep 27, 2026
I am hitting the limit on a free model on Go too, so it's not really free apparently
CinematicEnciclopedia commentedon Sep 28, 2026
Two new reports here since my last update, so a short status for anyone landing on this thread.
On 30/08 OpenCode Support told me by email that the block comes from a separate account-level monthly usage limit (in my case $40), and that "Use balance" does not override it. According to them, raising that limit in the workspace Billing settings lets requests fall back to the Zen balance. I haven't tested that myself: I waited for the monthly window to reset (~02/09) and it worked again. It might also explain the 401s mentioned above, although those could be a different problem.
So the fallback exists, but nothing in the UI says the toggle is capped by that limit, and the error doesn't say which limit you hit. That is what still needs fixing here: either "Use balance" should kick in once the Go quota is used, or the UI and the error should name the monthly limit and link to where it can be changed.
@khaosdoctor, thanks for opening #51682. The free-models block looks like the same provider-wide lock; I've added my data there.
thdxr commentedon Oct 11, 2026
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
CinematicEnciclopedia commentedon Oct 27, 2026
Update from OpenCode 2.0.29 (Windows). My original scenario (Go quota exhausted plus "Use balance") can't be tested any more: the workspace no longer has a Go subscription.
What I can confirm: requests through the standard endpoint do draw from the workspace credit. Sixteen Gemini 4 Argon requests on 16 Oct returned 200 and were billed.
What still hurts: requests for Grok 4.5/4.7 and GPT-6 Luna through
/inference/go/returned 403inference_failed(13–17 Oct), at no cost. The only message is "The inference request failed." I assume the missing Go subscription is the cause, but nothing says so. A clearer error would have saved a lot of debugging.I won't reopen this. Thanks for the pointer to 2.0.
khaosdoctor commentedon Oct 27, 2026
I dropped this, cancelled my subscription. If no one is taking care of it I won't pay for something that is broken