AI Relay 400, 429 and 503 Errors and Recovery Records
Understand AI relay HTTP 400, 404, 429, 503 and timeout evidence, with model- and protocol-specific failure counts and the next observed passing sample instead of provider-wide outage claims.
Check model and protocol for 400 or 404
HTTP 400 or 404 can reflect a model ID, request field, generation endpoint, or route compatibility mismatch. One sample is not treated as a provider-wide outage.
Separate rate, concurrency, and credit causes for 429
HTTP 429 can reflect a rate limit, concurrency limit, account credit, or route quota. It is not labeled insufficient balance without account-side evidence.
Retest the same conditions after 503 or timeout
A 503 or timeout means that request path did not complete. The next observed pass is recorded for the same provider, model, and protocol, but it is not presented as the exact recovery time.
Corrections
Providers can submit verifiable evidence or request a correction. Material changes retain a dated record.
Dated risk records
These entries separate observations, user reports, and conclusions. They do not use unverified fraud or exit labels, and website reachability is tracked separately from support and billing risk.
| Provider | Recorded | Signal | Evidence boundary |
|---|---|---|---|
| CCTQ | 2026-09-14 | Support route not verified | A user report dated September 14 said no usable support route was available. Our public check also did not find a clear official support, ticket, or after-sales entry. This is a support-risk record, not evidence that the API is unavailable. |
| 4Router | 2026-09-14 | Support route not verified | A user report dated September 14 said no usable support route was available. This is a support-risk report and does not establish API failure or billing misconduct. |
| Yes Code | 2026-09-14 | Support route not verified | A user report dated September 14 said no usable support route was available. This is a support-risk report and does not establish API failure or billing misconduct. |
| ByteCat | 2026-09-14 | Support route not verified | A user report dated September 14 said no usable support route was available. Verify support and refund channels before creating a key or adding prepaid balance. |
| Packy Code | 2026-09-14 | Support limitation reported | A user reported email-only support on September 14. Response time and after-sales handling have not been independently tested; this does not establish an API problem. |
| Micu | 2026-09-14 | Support limitation reported | A user reported that a community-access request was declined on September 14. The reason is unknown and the event does not establish an API failure. |
Before prepaying, confirm an official support channel, refund terms, account ownership rules, and whether a small test ticket receives a response. Corrections: airelaycheck@163.com
API failure and retest samples
Each row is a consecutive sequence in the recorded samples for one provider, model and protocol. The last column is the next observed passing request, not the actual recovery time. A single model failure does not establish a provider-wide outage. These are the 20 most recent sequences in the public per-provider evidence window, not a complete uptime history.
| Provider | Model / protocol | First / last failure (UTC) | Failed samples | HTTP / recorded error | Next observed pass (UTC) |
|---|---|---|---|---|---|
| APINebula # | claude-opus-5ClaudeCode | 2026-09-29T02:30:54.204Z2026-09-29T02:30:54.204Z | 1 | timeout | No later passing sample |
| OpenOx # | claude-fable-5-1OpenAI | 2026-09-18T10:15:56.845Z2026-09-29T02:29:58.738Z | 6 | HTTP 503 / http_503 | No later passing sample |
| H API # | gpt-5.6-terraResponses | 2026-09-18T13:18:29.161Z2026-09-29T02:29:51.767Z | 3 | HTTP 404 / http_404 | No later passing sample |
| UU API # | claude-opus-5Anthropic | 2026-09-25T02:29:58.973Z2026-09-25T02:29:58.973Z | 1 | HTTP 400 / http_400 | No later passing sample |
| H API # | gpt-5.5Responses | 2026-09-18T13:18:29.456Z2026-09-23T04:45:35.740Z | 4 | HTTP 404 / http_404 | No later passing sample |
| LinksAPI # | claude-fable-5Anthropic | 2026-09-18T10:16:36.569Z2026-09-22T02:30:38.439Z | 5 | timeout | 2026-09-27T02:30:05.790Z |
| UU API # | gpt-5.6-terraResponses | 2026-09-18T13:18:32.024Z2026-09-22T02:29:55.514Z | 4 | HTTP 404 / http_404 | 2026-09-27T02:30:09.129Z |
| boxying # | claude-fable-5ClaudeCode | 2026-09-18T10:15:50.354Z2026-09-20T10:09:35.742Z | 12 | HTTP 400 / http_400 | 2026-09-23T04:45:38.868Z |
| boxying # | claude-sonnet-5ClaudeCode | 2026-09-18T13:18:24.774Z2026-09-20T05:35:11.691Z | 11 | HTTP 400 / http_400 | 2026-09-22T02:29:55.427Z |
| boxying # | claude-opus-5ClaudeCode | 2026-09-18T13:18:24.136Z2026-09-20T03:33:19.371Z | 9 | HTTP 400 / http_400 | 2026-09-20T11:10:47.508Z |
| UU API # | claude-sonnet-5Anthropic | 2026-09-18T13:18:32.656Z2026-09-18T13:18:32.656Z | 1 | HTTP 400 / http_400 | 2026-09-18T14:39:59.333Z |
| UU API # | claude-opus-5Anthropic | 2026-09-18T13:18:32.112Z2026-09-18T13:18:32.112Z | 1 | HTTP 400 / http_400 | 2026-09-20T20:28:20.432Z |
| UU API # | claude-fable-5-1Anthropic | 2026-09-18T13:18:32.054Z2026-09-18T13:18:32.054Z | 1 | HTTP 400 / http_400 | 2026-09-19T16:42:30.277Z |
| UU API # | claude-fable-5-1Anthropic | 2026-09-18T10:15:59.262Z2026-09-18T10:15:59.262Z | 1 | HTTP 400 / http_400 | 2026-09-18T11:17:00.599Z |
Frequently asked questions
Does a risk record mean a relay is fraudulent?
No. A record describes a specific, dated observation or user report. It does not establish fraud, an exit, API failure, or future behavior unless separate evidence supports that conclusion.
Why is missing support treated as a risk?
Prepaid balances, refunds, key failures, and account restrictions may require support. If a usable official support or ticket route cannot be verified, users have fewer options when something goes wrong.
Can a provider correct a record?
Yes. Providers can send a verifiable official link, screenshot, status update, and effective date to airelaycheck@163.com. Material corrections retain a dated audit trail.
Same-model price and paid API evidence
Six provider profiles: compare input and output rates for the same model, then inspect the actual request outcomes. Live rankings and subscription estimates may differ from this build-time snapshot.
Claude Opus 5 · premium routes
| Provider | Price group | Input / 1M | Output / 1M | Source date | API passed / total | Latest test | P50 total | Best suited for |
|---|---|---|---|---|---|---|---|---|
| LinksAPI | Dedicated CC Max | ¥3.4 | ¥17 | 2026-09-25 | 4 / 4 | 2026-09-29 | 3.48 s | Users prioritizing a provider-labeled premium route |
| H API | Claude Max | ¥6.4 | ¥32 | 2026-09-25 | 3 / 3 | 2026-09-25 | 6.72 s | Claude Code or VS Code users who accept a client-only route |
| UU API | CC full-capability MAX - CC client only | ¥6.8 | ¥34 | 2026-09-25 | 1 / 3 | 2026-09-25 | 3.61 s | Claude Code or VS Code users who accept a client-only route |
| APINebula | Claude Code global protected route | ¥6.8 | ¥34 | 2026-09-25 | 3 / 4 | 2026-09-29 | 2.73 s | Claude Code or VS Code users who accept a client-only route |
| boxying | Claude official route | ¥8 | ¥40 | 2026-09-25 | 2 / 11 | 2026-09-25 | 2.66 s | Users prioritizing a provider-labeled premium route |
| OpenOx | direct | ¥19.2 | ¥96 | 2026-09-24 | 3 / 3 | 2026-09-27 | 3.81 s | Users prioritizing a provider-labeled premium route |
Prices are provider-published pay-as-you-go rates, not subscription equivalents or reconciled charges. API samples cover the named model, but the test key's actual upstream group is not consistently recorded and may differ from the quoted group. P50 is successful end-to-end response time, not time to first token. A completed call does not prove model identity.
GPT-5.6 Sol · premium routes
| Provider | Price group | Input / 1M | Output / 1M | Source date | API passed / total | Latest test | P50 total | Best suited for |
|---|---|---|---|---|---|---|---|---|
| H API | GPT Pro | ¥1.5 | ¥9 | 2026-09-25 | 3 / 3 | 2026-09-27 | 3.14 s | Codex or Responses API workloads after protocol testing |
| APINebula | CODEX | ¥1.95 | ¥11.7 | 2026-09-25 | No sample | - | - | Codex or Responses API workloads after protocol testing |
| UU API | Codex-GPT Pro account pool | ¥2 | ¥12 | 2026-09-25 | 3 / 3 | 2026-09-25 | 1.83 s | Codex or Responses API workloads after protocol testing |
| boxying | Codex premium pool | ¥2.2 | ¥13.2 | 2026-09-25 | 3 / 3 | 2026-09-25 | 10.8 s | Codex or Responses API workloads after protocol testing |
| LinksAPI | GPT Pro (Codex included), 0.3x | ¥3.25 | ¥26 | 2026-09-25 | 3 / 3 | 2026-09-27 | 17.0 s | Codex or Responses API workloads after protocol testing |
| OpenOx | Pro direct route | ¥9 | ¥54 | 2026-09-24 | 4 / 4 | 2026-09-29 | 2.73 s | Codex or Responses API workloads after protocol testing |
Prices are provider-published pay-as-you-go rates, not subscription equivalents or reconciled charges. API samples cover the named model, but the test key's actual upstream group is not consistently recorded and may differ from the quoted group. P50 is successful end-to-end response time, not time to first token. A completed call does not prove model identity.