OpenAI models leaving Cursor on November 12? What the official wording says (2026)
As of 2026-11-09, OpenAI calls November 12, 2026 a proposed transition period that Cursor may end sooner, and to keep GPT in Cursor it points to your own OpenAI API key (local Chat and Agent only, not Tab, Auto or Cloud Agents), the Codex IDE extension, or an AI gateway provider. The official termination date comes once both companies confirm it; Cursor's blog index and changelog list nothing on it yet, only a Cursor staff reply on its forum from 2026-08-29. A QCode key plugs in as an OpenAI-compatible gateway through Cursor's Override OpenAI Base URL field.
Updated 2026-11-10
Point this tool at the QCode endpoint — one key for Claude, GPT and Chinese models.
- Billed per token — live rates on /models
- Pay by card (Visa / Mastercard), Apple Pay, Google Pay or crypto
- Self-service: top up and activate instantly
Four things to know first
A proposed transition period, not a confirmed cutoff
OpenAI's help center proposes November 12, 2026 as a transition period during which it would keep providing Cursor access to the OpenAI models Cursor uses today. Cursor may choose to end access sooner, and the official termination date will be shared once both companies confirm it.
Your own key covers Chat and Agent, not the rest
OpenAI lists what an OpenAI API key does not cover in Cursor: Tab and autocomplete, Auto model selection and routing, Cloud or Background Agents, Automations, the Cursor CLI, and Cursor's API and SDK.
The Codex extension runs inside Cursor but does not change Cursor's models
OpenAI says the Codex IDE extension installs into Cursor and signs in with a ChatGPT subscription or an OpenAI API key. It is a separate extension and does not change the model used by Cursor Chat, Agent, Tab, Auto, Cloud Agents or other Cursor services.
Cursor's blog and changelog have no notice yet
On 2026-11-09 the Cursor blog index and the first page of its changelog (latest entry 2026-11-06) had no listed entry that mentions OpenAI. A Cursor staff reply on Cursor's forum dated 2026-08-29 said Cursor was speaking with OpenAI and would share more about continuity of access to OpenAI models.
What OpenAI and Cursor have actually said about November 12
As of 2026-11-09, OpenAI's help center article Using OpenAI models in Cursor says OpenAI is planning to wind down its contract providing OpenAI models to Cursor and has proposed a transition period of November 12, 2026. OpenAI's own statement of 2026-08-28 adds that it is giving the maximum notice its contract provides, that the agreement allows cancelling within a limited window after a change of control, and that it will not provide future models to Cursor. Cursor's own blog announced on 2026-08-14 that Cursor is now a part of SpaceX, and OpenAI's statement says it notified SpaceX of the wind-down. What can be planned on is therefore a contract for the OpenAI models Cursor uses today that is proposed to end on November 12 and may end earlier, with newer OpenAI models left out. On Cursor's side there is only a staff reply on its forum about the transition, and nothing about it in its blog index or changelog.
Latest check: still proposed, no new date
On 2026-11-09 the OpenAI help center still describes November 12 as a proposal and gives no other confirmed date. An article published on 2026-11-08 puts November 12 in its headline as the day OpenAI models leave Cursor, while its own text says November 12 is a proposed date, not a confirmed one. Treat the date as a working deadline that may come earlier, not as a promise of access until that day.
Timeline
OpenAI publishes a statement that it intends to wind down its contract providing OpenAI models to Cursor, with a proposed shutoff date of November 12, 2026, and will not provide future models to Cursor.
A Cursor staff member replies on Cursor's forum that Cursor is speaking with OpenAI, will share more about continuity of access to OpenAI models, and will continue to offer ways to access all models.
OpenAI's proposed shutoff date; on 2026-11-09 its help center still calls it a proposed transition period, and Cursor may end access sooner.
What is official and what is not settled
Confirmed (OpenAI and Cursor wording)
① OpenAI proposed a transition period of November 12, 2026 for the OpenAI models Cursor uses today. ② OpenAI will share the official termination date once both companies confirm it. ③ OpenAI says it will not provide future models to Cursor. ④ OpenAI's guide lists your own OpenAI API key (local Chat and Agent, billed at OpenAI API prices), the Codex IDE extension, and an AI gateway provider such as Amazon Bedrock, Azure, or another OpenAI-compatible gateway as ways to continue, and calls this guidance directional because Cursor might change functionality and availability over time. ⑤ Cursor's help page describes its OpenAI entry as standard, non-reasoning chat models and says Tab keeps using Cursor's built-in models. ⑥ A Cursor staff reply of 2026-08-29 said Cursor would share more about continuity of access to OpenAI models.
Not settled, so no guesses here
As of 2026-11-09 the sources above do not say which OpenAI models Cursor's picker will still list on and after November 12, whether your own key or a custom Base URL will be adjusted afterwards, or what the confirmed termination date is. A headline that states November 12 as a fixed cutoff goes further than OpenAI's proposed wording. Rely on OpenAI's help page and Cursor's own notices.
Ways to keep using GPT models for coding
Stay in Cursor: your own key and the Base URL override
In Cursor Settings → Models, put your key in the OpenAI entry; for an OpenAI-compatible endpoint also fill Override OpenAI Base URL. This covers local Chat and Agent only, not Tab, Auto or Cloud Agents. An account marked Cursor Staff on Cursor's forum said on 2026-08-21 that while the override is on, all OpenAI-family requests, including built-in picker models, go to your endpoint, and that the workaround is to turn it off when you use Cursor's own models. The related reading has a page on Cursor custom keys and Base URL.
Put GPT work next to Cursor: Codex
The Codex IDE extension installs into Cursor, signs in with ChatGPT or an OpenAI API key, has its own panel and does not change Cursor's models. Outside the editor, Codex CLI takes a custom provider in config.toml, and other tools with a custom Base URL setting are compared on the Cursor alternatives page in the related reading.
What to check before November 12
① Work out what you use OpenAI models for in Cursor: Chat and Agent can run on your own key, while Tab, Auto, Cloud or Background Agents, Automations and the Cursor CLI cannot. ② If you rely on built-in OpenAI entries, add your own key in Cursor Settings → Models now and test one request in Chat or Agent. ③ If you pay through a ChatGPT subscription, try the Codex IDE extension in Cursor and sign in with ChatGPT. ④ If you prefer token billing, point Override OpenAI Base URL at an OpenAI-compatible endpoint; the QCode setup is in the next section. ⑤ Read OpenAI's help page again after November 12, since the official termination date has not been announced.
Using GPT in Cursor through QCode
OpenAI's guide names an AI gateway provider such as Amazon Bedrock, Azure, or another OpenAI-compatible gateway as one way to keep OpenAI models in Cursor; QCode belongs to that group, as an OpenAI-compatible endpoint you enter in the Override OpenAI Base URL field. As of 2026-11-09, QCode's /models lists gpt-6.1-sol, gpt-6.1-terra, gpt-6-sol, gpt-6-astra, gpt-5.6-sol and gpt-5.6-terra, and one key covers Claude, GPT, GLM, Kimi, DeepSeek and Qwen, billed per token with per-model rates on /models. The QCode docs say to enter your cr_ key and turn on Override OpenAI Base URL with https://api.qcode.cc/openai/v1; like any gateway credential it serves Cursor's local Chat and Agent, requests still pass through Cursor's servers, and OpenAI notes that some model-specific controls may not be available or forwarded through a custom base URL. Three steps: ① register and top up (Alipay, WeChat Pay, cards, crypto and more); ② buy a Personal Plan or a Pay-as-you-go Pack and get your API key, which starts with cr_; ③ enter it in Cursor as the QCode docs describe, or set up Codex CLI. Not ready for a monthly plan? Pay-as-you-go Packs start at $4 for a one-time quota you use within 7 days; if you already have a Personal Plan, add a booster pack when you need more quota.
Frequently asked questions
Can I still use GPT in Cursor after November 12?
OpenAI's guide says yes through your own OpenAI key (local Chat and Agent), the Codex IDE extension or a compatible AI gateway; what OpenAI proposes to end on November 12 is its contract providing OpenAI models to Cursor. OpenAI calls this guidance directional and says Cursor might change functionality and availability, so check again after the date.
Does Cursor BYOK work with Tab?
No. OpenAI says Cursor doesn't support bring your own key in Cursor Tab, Auto, Cloud or Background Agents, Automations, the Cursor CLI, or Cursor's API and SDK, and Cursor's help page says Tab completion continues using Cursor's built-in models.
Is November 12 a confirmed shutoff date?
No. OpenAI's wording is a proposed transition period of November 12, 2026, Cursor may choose to end access sooner, and OpenAI says it will share the official termination date once both companies confirm it. Some headlines state November 12 without the word proposed.
Will my own OpenAI API key or custom Base URL stop working?
None of the pages checked on 2026-11-09 says it will. OpenAI lists your own key and a custom OpenAI-compatible Base URL as ways to continue, with the caveat that Cursor might change these features over time. Which OpenAI models you can enable depends on what Cursor's model picker lists and on your OpenAI API organization.
Does the Codex extension replace Cursor's OpenAI models?
No. It gives you the Codex agent inside Cursor, not the model picker: OpenAI says it runs as a separate extension, signs in with a ChatGPT subscription or an OpenAI API key, and does not change the model used by Cursor Chat, Agent, Tab, Auto, Cloud Agents or other Cursor services.
Can I use GPT in Cursor through QCode?
Yes. QCode's Cursor guide uses the OpenAI protocol: enter your cr_ key, turn on Override OpenAI Base URL with https://api.qcode.cc/openai/v1, then add a model that /models lists, for example gpt-6.1-sol; requests still go through Cursor's servers. For Codex CLI the docs use base_url = https://api.qcode.cc/openai with wire_api = responses. Billing is per token, with per-model rates on /models.
Sources
OpenAI: the help center article Using OpenAI models in Cursor and its statement on the Cursor decision (news feed date 2026-08-28), read on 2026-11-09. Cursor: the help center page Bring your own API key, the blog index, the first page of the changelog, and two forum threads with replies from accounts marked Cursor Staff (2026-08-21 and 2026-08-29), read on 2026-11-09. QCode: qcode.cc/models and the Cursor and Codex pages on docs.qcode.cc, retrieved on 2026-11-09. One news article of 2026-11-08 is cited for its headline and for its own note that the date is proposed.
Keep GPT models in your coding workflow
One QCode key works for GPT models in Cursor's OpenAI entry and in Codex CLI, billed per token; per-model rates are on /models.
Related reading
Pay-as-you-go Packs
One-time quota packs valid for 7 days, no subscription; per-model rates are on /models.
Cursor custom API key and Base URL: where to enter them
Which Settings → Models entry takes a Base URL, why Tab skips your key, and the QCode setup.
Cursor alternatives 2026: switch IDE, use a terminal agent or add Chinese models
Other tools that take a custom Base URL, with versions checked on 2026-11-05 and the QCode setup.
Crypto payment guide
Nine coins, 56 networks, zero platform fee for AI coding API credits.
Statements on this page were checked on 2026-11-09 against OpenAI's help center and statement and Cursor's own pages, and the official pages prevail. What you can use in Cursor depends on your Cursor setup and plan; QCode's current models and per-model rates are on /models.