gpt-oss reasoning /thinking stream format not detected

[dirkpetersen]

I am running gpt-oss-120b as a custom on-premises endpoint, the performance is awesome, however librechat does not not seem to detect the reasoning format and the "Thinking" widget is not visible.

I am running the latest librechat dev version as of today and set openai format:

  custom:
    - name: "XXXX-XXX"
      apiKey: "${DUMMY_KEY}"
      baseURL: "http://dgx01.o.edu:8000/v1"
      models:
        default: ["openai/gpt-oss-120b", "openai/gpt-oss-20b"]
        fetch: true
      titleConvo: true
      titleModel: "current_model"
      summarize: false
      summaryModel: "current_model"
      forcePrompt: false
      modelDisplayLabel: "openAI GPT OSS"
      customParams:
        defaultParamsEndpoint: 'openAI'
        paramDefinitions:
          - key: temperature
            range:
              min: 0
              max: 1.0
              step: 0.1
            default: 0.7

I see this :

and have looked at #9053 and #9085 but they seem to solve a different issue ?