OpenAI Integration does not work with unverified accounts

[XHyperDEVX]

Describe the feature

The option to disable “Reasoning Summaries”

Use cases

This is necessary in order to use GPT5 models with an unverified OpenAI account, as OpenAI does not allow reasoning summaries via API keys from unverified OpenAI accounts.
Currently, it is not possible to use GPT5 models in Home Assistant with an unverified OpenAI account, so this is urgently important.

Anything else?

Home Assistant Log:

ERROR (MainThread) [homeassistant.components.openai_conversation] Error talking to OpenAI: Error code: 400 - {'error': {'message': 'Your organization must be verified to generate reasoning summaries. Please go to: https://platform.openai.com/settings/organization/general and click on Verify Organization. If you just verified, it can take up to 15 minutes for access to propagate.', 'type': 'invalid_request_error', 'param': 'reasoning.summary', 'code': 'unsupported_value'}}

[mik-laj]

Out of curiosity - is there a specific reason you’re not able (or don’t want) to verify your OpenAI account? Just wondering if there’s some limitation or friction in that process we should be aware of.

From our side, I’m not entirely sure that supporting unverified accounts is something we’d want to invest effort in. The OpenAI API itself restricts reasoning-capable models to verified organizations, and building workarounds for that would make testing and maintenance harder while benefiting only a small number of users.

It might make more sense to document this as a known limitation rather than trying to implement support for unverified accounts.

[XHyperDEVX]

@mik-laj
Yes, the same reason why I use Home Assistant and not just any proprietary smart home system. Anonymity and privacy.

There is no problem with OpenAI verification itself, but OpenAI wants to permanently store photo ID/drivers license including a picture of yourself. This contradicts the Home Assistant Mindset on all levels.
In the past, such databases have often been leaked (recently it was "Discord") and the oh-so-securely stored "sensitive data" has been stolen and is sometimes used for identity theft.

And for obvious reasons, I don't want to risk that, especially for a feature that I wouldn't use in Home Assistant anyway.

I understand your concern about allowing unverified accounts, but let me briefly explain what OpenAI verification is supposed to achieve.
It is NOT verification like confirming your email after creating an account. The point is that if OpenAI determines that you are training other AI models with your account (which violates OpenAI guidelines), they can contact the authorities directly with all your data. You can legally use the OpenAI API without a verified account, but with restrictions. (For example, no "reasoning summary," which is the problem here.)

Furthermore, adding this feature would not be a "workaround" to circumvent OpenAI's policies. Unverified accounts are tolerated by OpenAI; they are only restricted. Adding this feature would simply be adding a function to Home Assistant from the official OpenAI API.
It is not immoral; on the contrary, it would free Home Assistant from having to verify itself with OpenAI in order to use the integration with the latest (and cheapest) models.

If you have any further questions, I will be happy to answer them.
I would be delighted to see this feature integrated soon, as the effort involved should be very limited.