GPT5 not working with ResponseStreaming

[cescofry]

Describe the bug When setting .gtp5 in the query for response streaming I get a 400.

Version 0.4.6

Setup

let query = CreateModelResponseQuery(
            input: .textInput("Tell me a joke"),
            model: .gpt5,
            stream: true
        )
        
        let stream = client.responses.createResponseStreaming(query: query)

Error Headers

{
    "Alt-Svc" =     (
        "h3=\":443\"; ma=86400"
    );
    "CF-RAY" =     (
        "xxxxxxxxxxx-SJC"
    );
    Connection =     (
        "keep-alive"
    );
    "Content-Length" =     (
        191
    );
    "Content-Type" =     (
        "application/json"
    );
    Date =     (
        "Wed, 13 Aug 2025 22:27:17 GMT"
    );
    Server =     (
        cloudflare
    );
    "Set-Cookie" =     (
        "_cfuvid=xxxxxxxxx_xxxxxxxx-1755124037886-0.0.1.1-604800000; path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None"
    );
    "Strict-Transport-Security" =     (
        "max-age=31536000; includeSubDomains; preload"
    );
    "X-Content-Type-Options" =     (
        nosniff
    );
    "cf-cache-status" =     (
        DYNAMIC
    );
    "openai-organization" =     (
        "user-xxxxxxxxx"
    );
    "openai-processing-ms" =     (
        25
    );
    "openai-project" =     (
        "proj_xxxxxxxxx"
    );
    "openai-version" =     (
        "2020-10-01"
    );
    "x-envoy-upstream-service-time" =     (
        29
    );
    "x-request-id" =     (
        "req_xxxxxxxxx"
    );
}

Smartphone:

• Device: iPhone 13 mini, and iOS Simulator
• OS: iOS18.6.1
• Version 0.4.6

[cescofry]

[UPDATE] Done some investigation and when running the query using CURL I get the error:

Your organization must be verified to stream this model. 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

curl https://api.openai.com/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <YOUR_TOKEN>" \
  -d '{
    "model": "gpt-5",
    "input": "Write a haiku about code.",
    "stream" : true,
    "reasoning": { "effort": "low" }
  }'

I did verified my organization (personal account), but while this fixed the CURL call, I still see the same error using the library.

[cescofry]

[UPDATE] The remaining problem was that GTP5 does not accept temperature as input. However I could only see the issue by using a tool like Charles or ProxyMan, and inspect the response body:

{
  "error": {
    "message": "Unsupported parameter: 'temperature' is not supported with this model.",
    "type": "invalid_request_error",
    "param": "temperature",
    "code": null
  }
}

I guess this should be the error passed to ResponseEndpoint.createResponseStreaming

[mi12-root]

It would be great if we improve the error API.