For model not found or /v1/v1, first check the model ID and final request URL, then permissions, catalog changes and upstream availability. Repeated top-ups, key creation or retries usually do not identify the cause.
Quick conclusion
Check the URL actually sent by the client, the HTTP status and error body, then use the same key to query /v1/models and copy the complete id. If you see /v1/v1 fix the path first. For model not found in JSON, check the model ID, endpoint and permissions. An HTML 404 may come from a frontend fallback, proxy or error page; it does not by itself prove that a model was removed.
Fix model not found: follow the response type
| Observed response | Next step | What this does not establish |
|---|---|---|
| JSON reports model not found or model_not_found | Use the same key to check the model ID, token permissions and target endpoint | That every model on the platform is unavailable |
| HTML 404 or webpage content | Check the actual host, path, Content-Type and proxy error page | That the request reached the upstream model |
| The final path contains /v1/v1 | Remove the version path duplicated by the client and configuration | That the API key must be wrong |
| /v1/models returns 401 | Resolve authentication before troubleshooting the model | That the model was removed |
| The target ID is absent from the model list | Check account visibility and whether the catalog is complete | That no account can use the model |
| The target ID is listed but generation fails | Check the ID actually sent, endpoint type and error body | That successful listing guarantees generation |
The HTTP 404 definition also allows a service to conceal whether a resource exists. Use the response body and permission checks rather than one status code to infer platform capabilities.
Check before you start
Log the full URL and model fields that the client actually sent. Don't just look at what is filled in the settings page, because the client may automatically append paths or replace model aliases.
When you are not sure how the addresses will be spliced, you can use Base URL online checkerGenerate recommended values and final request path. This tool only analyzes the URL string in the browser, does not request the target API, and does not require an API Key.
Configuration steps
Get the real model ID
if [ -n "$AIFAST_API_KEY" ]; then
curl --connect-timeout 10 --max-time 30 -sS -i \
https://www.aifast.hk/v1/models \
-H "Authorization: Bearer $AIFAST_API_KEY"
else
printf '%s\n' 'AIFAST_API_KEY is not set'
fiUse only a key created on HK, and copy the complete id field from the JSON response, preserving case and suffixes. The following usually cannot be used directly as model IDs:
- Web card title.
- Chinese model name.
- Manufacturer name.
- Marketing labels such as “latest”, “flagship” and “high-speed”.
Check final URL
A correct Chat Completions request usually looks like:
https://www.aifast.hk/v1/chat/completionsError example:
https://www.aifast.hk/v1/v1/chat/completions
https://www.aifast.hk/chat/completions
https://www.aifast.hk/v1/models/chat/completionsIf the client automatically appends /v1, Base URL fills in the root domain name; if the client requires a complete compatible address, keep it /v1.
FAQ
It is in the model list, but the call still does not exist
A model may support only certain endpoints, or token and group permissions may differ. Successful listing does not guarantee support for Chat Completions, Responses, image and video endpoints. Keep the same key and model ID, then follow the provider documentation to compare short requests while changing only the endpoint. Generation requests may incur charges.
404 returns the web page HTML
The request may have reached a frontend or proxy error page. Check the host, path, Content-Type and request ID; do not treat an HTML error page as a model response.
What to include in a support request
Include the timestamp and time zone, client version, redacted full path, endpoint type, model ID actually sent, HTTP status, error body and request ID. Do not share complete API keys, cookies, authentication headers or production prompts. Without generation requests and server logs, configuration checks alone cannot establish that a model has recovered.
The tool automatically changes the model name
Some clients only allow built-in models or map display names to fixed IDs. Check the request log to confirm the final sent value; use the custom model function provided by the client if necessary.
Next step
After correcting the model name and address, use non-streaming short request verification. After the short request succeeds, tool calls, long context, and streaming output are resumed.
If it still returns 401, 429, or 5xx, continue checking API error troubleshooting; once the URL is correct, runModel Quality Checkto inspect the protocol and response fields. Results alone cannot prove the underlying model identity, and API calls may incur charges.
When converting a project using a community protocol such as Cursor2API, the model directory may also change with the upstream entry. Please combine Cursor2API Risk Check and Migration GuideDetermine whether the model ID configuration is incorrect or the upstream model has been adjusted.