Common Errors and Solutions
Model-related Issues
Q: Why is there no output after making a request?
Possible reasons include:
- The client disconnected just before output.
- Tools call or function call was triggered.
[DONE]was returned, generally due to policy or security reasons refusing to answer. Check the returnedfinish_reasonfor diagnosis.
Q: "No available channels under current group" prompt?
Please ensure the model name matches the list exactly and is case-sensitive.
Q: Why is GPT-4 quota consumed so quickly?
GPT-4's consumption rate is 20 to 40 times that of GPT-3.5. With historical messages attached to each request, the number of messages you can send further decreases. Please use with caution.
Q: Invalid token?
This issue may be caused by:
- Incorrect address, not our API address.
- Token not set properly in the program.
- Token has expired, please generate a new token.
Quota-related Issues
Q: Backend quota is sufficient, but API returns insufficient quota?
Please confirm that the token you created in the backend has been allocated quota. The expiration time can generally be set to unlimited. Also, quota is not filled in as an amount; 500,000 quota = 1 USD. Fill in as needed.
Q: What is quota? How is it calculated?
Quota = Group multiplier × Model multiplier × (Prompt tokens + Completion tokens × Completion multiplier)
The completion multiplier is fixed at 1.33 for GPT-3.5 and 2 for GPT-4, consistent with official settings.
Q: "Quota for this token has been exhausted"?
This means your token is out of quota. You need to add quota to your token, or your wallet is out of quota and you need to recharge.
Q: "user quota is not enough"?
This means your account is out of quota and you need to recharge.
Network-related
Q: Failed to fetch?
Your network environment may be problematic. Please try switching networks. Or you may have used the wrong API address, please check if the API address is correct.
Q: CDN origin error occurs?
Please contact customer service to get the correct API address.
Other Issues
Q: max_tokens is too large?
The max_tokens parameter in your request is set too high. The total tokens in context + max_tokens must be less than or equal to the model's context size.
Q: Some website pages report errors upon opening?
Please clear your browser cache and cookies, then refresh the page.
Common HTTP Error Codes
| Error Code | Description |
|---|---|
| 400 Bad Request | Invalid or malformed request format, check request parameters |
| 401 Unauthorized | Invalid request token, check API Key |
| 403 Forbidden | Insufficient balance or restricted access permissions |
| 404 Not Found | Requested resource not found, check endpoint address |
| 413 Request Entity Too Large | Request body too large, reduce data payload |
| 429 Too Many Requests | Exceeded rate limit, lower request frequency |
| 500 Internal Server Error | Internal server error |
| 503 Service Unavailable | Service temporarily unavailable, possibly in maintenance |