FAQ: Why is my request too large?

request_too_large means the single request is estimated to exceed the gateway's per-request token capacity.

This is not a temporary rate limit. Retrying the same payload will not succeed.

Fix it

• Remove unnecessary context.
• Split the task into smaller requests.
• Summarize previous conversation before sending it again.
• Reduce attached or pasted data.

If the task genuinely needs larger context, contact support with the model ID, estimated input size, and use case.