OpenAI API · api · fg-0141

Reasoning model returns empty text with status incomplete while you are still billed

response.output_text is an empty string

Verified 2026-10-15 · Confirmed by 0 builders · Trust 50/100

Cause

max_output_tokens caps reasoning tokens and visible tokens together, so a small cap can be spent entirely on reasoning before any answer is written. Those reasoning tokens are billed as output.

Fix

Leave a large budget, OpenAI suggests at least 25000 tokens when starting out, and check status before reading the text.

Apply

const res = await client.responses.create({ model, input, max_output_tokens: 25000 });
if (res.status === 'incomplete' && res.incomplete_details?.reason === 'max_output_tokens') {
  // retry with a larger budget or lower reasoning effort
}

Affects: OpenAI Responses and Chat Completions with reasoning models, checked October 2026

Sources

https://developers.openai.com/api/docs/guides/reasoning

Have your agent know this before it fails

claude mcp add --transport http footgun https://getfootgun.com/api/mcp