Hey @bejmorri! Yeah, I can see how that would be frustrating, especially with requests sitting queued for 5–10+ minutes.

What @sps shared is the right place to track this. There was a GPT5.5 performance degradation incident reported on May 15, mitigation was applied on May 16, and the status page now shows all impacted services have fully recovered.

For future updates on service issues, you can…

@OpenAI_Support @sps
I’m hitting what looks like a silent failure on the Realtime API’s REFER endpoint. The HTTP request succeeds, but no SIP REFER message ever leaves OpenAI’s SIP backend — the call stays on the original audio path and the transfer never happens.

Setup

• Model: gpt-realtime-1.5

• Inbound SIP carrier: Twilio Elastic SIP Trunk → sip.api.openai.com;transport=tls

*…

i remember this announcement way back when, but @sps mentioned today they’re finally launching some interesting tech…

Here are the key details regarding the launch and company status as of May 12, 2026:

• Core Product: “Interaction Models”

  • Thinking Machines is launching a new class of models, specifically TML-Interaction-Small , designed for native, real-time…

I’m not sure what you’re saying here.

I’m saying that the string “The user is in an estimated location of United States” is injected into the prompt when tools=[{"type": "web_search"}]is passed to client.responses.create.

I’m not talking about what happens when a location is passed in the web search definition, or how the model infers location from language, or the scope of search results,…

Hi dear @sps

Thanks for your reply - Yes and we do use this approach.

And then how can I support language switching? Example

1. Customer is in the US, and greets you in English.
2. Caller is Spanish and starts in Spanish.
3. It’s currently very convenient when the realtime turn gets to be in Spanish. Excellent negotiation.

How can I have support for both?

@Abrie_Cronje @sps Have you found a solution? I am having the same issue unfortunately.

Hello @sps !

The article you shared points to platform.openai .com → data controls → sharing (tab). All three of the following options stated in the sharing (tab/page) are disabled for me:

• Enable sharing of model feedback from the Platform
• Share evaluation and fine-tuning data with OpenAI
• Share inputs and outputs with OpenAI

@sps thanks for the suggestion.

We already use tools for other things and could add a deterministic calendar tool. The challenge is that inputs aren’t always clean YYYY-MM-DD dates; sometimes callers just say a weekday (e.g. “next Tuesday” or “this Friday”), or even provide both.

One approach could be a single flexible tool with optional arguments and let the model decide how to use it. Then…

I am facing the same issue, if someone finds a solution kindly share. @sps