Go free models documented as "Unlimited" are blocked when any Go usage cap is reached

[khaosdoctor]

When a Go usage limit is reached (5h, weekly, or monthly), the block applies to the entire opencode-go provider. This includes the free models that are documented as Unlimited.

The Go docs list Space Bunny Free and LongCat 2.5 Preview Free with a monthly limit of Unlimited (limited time). These models have no dollar cap and should not be subject to any usage limit. However, when any Go model's cap is hit, requests to these free models are blocked with the same "5 hour usage limit reached" error.

Expected behavior

Free models on Go remain usable when paid-model caps are reached, since they carry no dollar limit. The block should apply only to the specific model whose cap was hit.

Actual behavior

All Go models, including free ones, become unusable. The only workaround is switching to the Zen provider and using free models there, which defeats the purpose of the Go subscription.

Steps to reproduce

1. Subscribe to OpenCode Go
2. Use any Go model until a usage limit is reached (5h, weekly, or monthly)
3. Attempt to use opencode-go/space-bunny-free or opencode-go/longcat-2.5-preview-free
4. Observe the same "5 hour usage limit reached" error

Environment

• OpenCode version: 1.18.31
• OS: macOS
• Terminal: macOS Terminal

Related issues

• Go plan hits 100%, blocks 12h — "Use balance" enabled but $39.89 Zen balance never used, on any model #42938: "Use balance" enabled but Zen balance never drawn when Go cap is reached
• Free usage exceeded and there's long retry timers keep escalating across different free models #50093: Retry timers escalate across free models instead of counting down
• Go subscription active but still "Free usage exceeded" — paid, no access #36265: Go subscription active but all models blocked

[github-actions]

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

• #49014: A Go model reaching its 5-hour limit blocks other models, despite their individual usage being below limit.
• #47703: Free models are also blocked after Go usage limits are reached.
• #43899: Go free models stop working after a usage limit is hit.

[khaosdoctor]

Slightly different thing, it includes all the other issues

[CinematicEnciclopedia]

Confirming from #42938: when my Go quota hit 100%, all models on the Go provider were blocked (~12h), so it looks like the same provider-wide lock you describe here.

For context: OpenCode Support told me by email (30/08) that the block I hit came from a separate account-level monthly usage limit, which "Use balance" does not override. That explains why the Zen balance isn't drawn, but not why models documented as Unlimited are locked as well. Agree that the lock should apply per model (or at least leave the free ones out), and that the error should say which limit was reached.

Cross-referencing so both get triaged together: #42938

[shimpeiws]

I encountered a similar issue and found an additional workaround:

1. I reached 100% of my Go monthly usage limit.
2. A request using a free model was also blocked with a monthly limit message.
3. I added $5 in credits, but requests were still blocked while Extra Usage was disabled.
4. After enabling Extra Usage, I switched back to the free model, and it worked.

So, in my case, enabling Extra Usage allowed the free model to work again without switching to the Zen provider.

I haven’t verified whether the successful free-model requests deducted any credits. However, this suggests that access to free models may also depend on the Extra Usage setting once the Go quota is exhausted.

[chrisssp]

I am experiencing this exact issue. I have the Go plan ($10/month) with 43% of the monthly limit remaining, and my 5-hour and weekly limits are at $0 (completely unused), yet I get "Go usage limit exceeded" on all models — including free models like mimo-v2.6-flash-free, ling-3.1-flash-free, and big-pickle.

The logs show mixed errors: "Go usage limit exceeded", "ECONNRESET", "Endpoint is unavailable", and "Rate limit exceeded", all in bursts of ~13 minutes.

My config uses opencode-go/mimo-v2.5 for build/plan/general and opencode-go/deepseek-v4.1-flash for explore. If one model's 5-hour limit blocks all others, this makes the plan unusable for real work.

Please prioritize the fix — the docs promise per-model limits but the implementation blocks the entire account. Thanks.