431 Request headers are too large

API Bugs
,
1

It happens rarely. When running Chat Completion API requests.
library: openai v6.39.1
node.js v22.22.2

model: gpt-4.1
subject: ワー…した (maybe it is related?)
message: ‘Request headers are too large.’,
code: ‘request_headers_too_large’

Response Headers
{ date: ‘Thu, 11 Jun 2026 22:15:31 GMT’, ‘content-type’: ‘application/json; charset=utf-8’, ‘content-length’: ‘186’, connection: ‘keep-alive’, ‘cf-ray’: ‘a0a3f992eca2c99d-IAD’, ‘cf-cache-status’: ‘DYNAMIC’, server: ‘cloudflare’, ‘strict-transport-security’: ‘max-age=31536000; includeSubDomains; preload’, vary: ‘Origin’, ‘access-control-expose-headers’: ‘CF-Ray, CF-Ray’, ‘x-content-type-options’: ‘nosniff’, ‘x-openai-proxy-wasm’: ‘v0.1’, ‘x-ratelimit-limit-requests’: ‘10000’, ‘x-ratelimit-limit-tokens’: ‘30000000’, ‘x-ratelimit-remaining-requests’: ‘9999’, ‘x-ratelimit-remaining-tokens’: ‘29999022’, ‘x-ratelimit-reset-requests’: ‘6ms’, ‘x-ratelimit-reset-tokens’: ‘1ms’, ‘x-request-id’: ‘req_f39702dcae9f4309904798df1de9c170’, ‘set-cookie’: ‘__cf_bm=yaJUK2oms6x6MUlY4DCbq_Sy2rDesPuogNcmmPV.xdA-1781216131.0301845-1.0.1.1-CSy.dFILQzyhacGNZcrFDl1mZF6CQsZk9InVftmb0draRReBK1szkh13tXmEoxbSvZClEouNXXXOuRQPZm3WrlTQCFEncMBDdI3NXX8.FCGTB3rchmqLS3xixeXgTelP; HttpOnly; SameSite=None; Secure; Path=/; Domain=api.openai.com; Expires=Thu, 11 Jun 2026 22:45:31 GMT’, ‘alt-svc’: ‘h3=\“:443\”; ma=86400’ },\n requestID: ‘req_f39702dcae9f4309904798df1de9c170’,\n error: { message: ‘Request headers are too large.’, type: ‘invalid_request_error’, param: null, code: ‘request_headers_too_large’ },\n code: ‘request_headers_too_large’,\n param: null,\n type: ‘invalid_request_error’ }

2

I constantly experience these OpenAI errors:
An error occurred while processing your request. You can retry your request, or contact us through our help center at help.openai.com if the error persists. Please include the request ID req_60a6ebbfc0ab4114ac9eb498ae70a1b6 in your message.

I already opened 3 tickets (10056226, 09753963, 09428643) to OpenAI support but their support was absolutely indifferent and unprofessional.
As I see, they are more used to see issues occurred in their UI, not by the API.