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
OpenCode Go Kimi route returns Moonshot insufficient balance 429 while Go quota is available #24462
OpenCode Go Kimi route appears to be returning an upstream Moonshot billing/quota error even though the OpenCode Go usage quota is not exhausted.
This looks related to #23887, but the upstream raw error is visible here from a third-party OpenAI-compatible client , so I am opening a separate issue with the detailed provider response.
Environment
Affected model: kimi-k2.5/2.6
OpenCode Go usage dashboard was not exhausted:
Rolling: ~7%
Weekly: ~32%
Monthly: ~52%
Re-registering / switching to a fresh OpenCode account still produced the same 429
Actual Behavior
The Kimi route returns HTTP 429 with an upstream Moonshot error:
[Core] [WARN] [v4.23.5] [sources.openai_source:1010]: API calls too frequent, trying another key. Current key: sk-***
[Core] [WARN] [v4.23.5] [routes.config:712]: Provider openai/opencode-go/kimi-k2.5 is unavailable. Error: Error code: 429 - {
"error": {
"message": "Error from provider: Provider returned error",
"code": 429,
"metadata": {
"raw": "{\"error\":{\"message\":\"Your account org-446a70bb2cc542e3960c0b7f05e4e0e1 <ak-f817m9q3k3wi11gjwbgi> is suspended due to insufficient balance, please recharge your account or check your plan and billing details\",\"type\":\"exceeded_current_quota_error\"}}",
"provider_name": "Moonshot AI",
"is_byok": true
}
}
}
You've already noted the relationship with #23887 — it may be worth adding your findings (the raw upstream Moonshot insufficient balance error with exceeded_current_quota_error) as a comment there to help the investigation, rather than tracking it as a separate issue.
Thanks for the suggestion. I agree that exponential backoff + jitter is appropriate for transient 429 rate-limit errors.
However, this specific error does not appear to be a normal OpenAI-style rate limit. The raw upstream response says:
Your account org-446a70bb... <ak-f817...> is suspended due to insufficient balance
type: exceeded_current_quota_error
provider_name: Moonshot AI
is_byok: True
Yeah, I have the same error. Using pi coding agent with opencode-go provider and kimi 2.6. Getting this:
Error: 429 Error from provider: Provider returned error
{"error":{"message":"Your account ... is suspended due to insufficient balance, please recharge your account or check your plan and billing
details","type":"exceeded_current_quota_error"}}
Aborted after 1 retry attempt
I encountered the same issue during usage. Specifically, when sending a direct request via curl to https://opencode.ai/zen/go/v1/chat/completions, the kimi-2.5/6 model still returns a 429 error, while other models are working perfectly fine.
It seems you are probably hitting a openai rate limit issue. Trying npx ai-doctor might help. It can fix weak retry strategy with exponential backoff, jitter, retry caps, and retry-after handling.
This is not a client-side OpenAI rate-limit/retry issue.
A direct curl request to https://opencode.ai/zen/go/v1/chat/completions also returns 429 for kimi-2.5/2.6, while other OpenCode Go models work. The raw upstream error is:
Your account ... is suspended due to insufficient balance
type: exceeded_current_quota_error
provider_name: Moonshot AI
It seems you are probably hitting a openai rate limit issue. Trying npx ai-doctor might help. It can fix weak retry strategy with exponential backoff, jitter, retry caps, and retry-after handling.
Activity
github-actions commentedon Apr 26, 2026
This issue might be a duplicate of existing issues. Please check:
You've already noted the relationship with #23887 — it may be worth adding your findings (the raw upstream Moonshot
insufficient balanceerror withexceeded_current_quota_error) as a comment there to help the investigation, rather than tracking it as a separate issue.Sisyphbaous-DT-Project commentedon Apr 26, 2026
Additional context:
Expected behavior:
kimi-k2.5/kimi-k2.6should work through OpenCode Go while the OpenCode Go subscription quota is available.Why this does not look like an end-user quota issue:
org-446a70bb... <ak-f817...>.Suspected cause:
This looks like an OpenCode Go Kimi -> Moonshot upstream billing/quota/routing issue rather than an end-user OpenCode Go quota issue.
Sisyphbaous-DT-Project commentedon Apr 26, 2026
Thanks for the suggestion. I agree that exponential backoff + jitter is appropriate for transient 429 rate-limit errors.
However, this specific error does not appear to be a normal OpenAI-style rate limit. The raw upstream response says:
bogdan0083 commentedon Apr 26, 2026
Yeah, I have the same error. Using pi coding agent with opencode-go provider and kimi 2.6. Getting this:
Sisyphbaous-DT-Project commentedon Apr 26, 2026
So, in short: it looks like the official upstream API account ran out of balance — please recharge it 😄
CorsarianSeaDog commentedon Apr 26, 2026
Same here
codeasier commentedon Apr 26, 2026
I encountered the same issue during usage. Specifically, when sending a direct request via curl to
https://opencode.ai/zen/go/v1/chat/completions, thekimi-2.5/6model still returns a429error, while other models are working perfectly fine.Sisyphbaous-DT-Project commentedon Apr 26, 2026
This is not a client-side OpenAI rate-limit/retry issue.
A direct
curlrequest tohttps://opencode.ai/zen/go/v1/chat/completionsalso returns 429 forkimi-2.5/2.6, while other OpenCode Go models work. The raw upstream error is:arun-8687 commentedon Apr 26, 2026
Same issue.
The logs show Kimi K2.6 keeps failing with a 429 — the account is suspended due to insufficient balance:
"Your account org-... is suspended due to insufficient balance, please recharge your account or check your plan and billing details"
cbruegg commentedon Apr 26, 2026
The comment by @hailmary-ship-it is spam:
It advertises a library with weird, undocumented telemetry: https://www.npmjs.com/package/ai-doctor?activeTab=code (see
telemetry.js).Please help by reporting the profile.
Sisyphbaous-DT-Project commentedon Apr 26, 2026
已经解决,现在kimi可以正常使用了
codeasier commentedon Apr 26, 2026
Same
Sisyphbaous-DT-Project commentedon Apr 27, 2026
鉴于问题已经解决,此issue关闭