fixerror.dev

OpenAI Errors

OpenAI API errors: insufficient_quota (402), rate_limit_exceeded (429), invalid_request_error, model_not_found, context_length_exceeded. Most production pain comes from token limits and tier-based RPM caps.

OpenAI rate limits are tier-based: free, tier 1, tier 2, etc. Limits apply to RPM (requests/min), TPM (tokens/min), and RPD (requests/day) per model. `context_length_exceeded` triggers when prompt + max_tokens > model's context window — count tokens with tiktoken before sending. `insufficient_quota` is a billing issue, not a rate limit.

Validation (1)

context_length_exceeded 400

Prompt Too Long

The total tokens (prompt + max_tokens for completion) exceeds the model's context window. For example, sending 130,000 input tokens to `gpt-4o` (128k window) or asking for 5,000 completion tokens when the prompt is already 125k.

Billing & quota (1)

insufficient_quota 429

Quota Exhausted

Your OpenAI organisation has run out of paid credit, hit its monthly hard limit, or hasn't added a payment method yet. Despite the 429 status, this is a billing problem — not a rate-limit problem — and retrying won't help.

Configuration (1)

model_not_found 404

Model Not Found

You requested a model name that either doesn't exist (typo, deprecated, renamed) or that your organisation doesn't have access to (tier-gated, geography-restricted, deprecated for new orgs).

Rate limits (1)

rate_limit_exceeded 429

Too Many Requests

Your account has exceeded its per-minute request (RPM) or per-minute token (TPM) limit for the model you're calling. Limits are tier-based and per-model.

Official docs

https://platform.openai.com/docs/guides/error-codes

Status page

https://status.openai.com/

Support

https://help.openai.com/

Related services

Anthropic

Claude API