Models supported by the Responses API The Responses API currently only supports the DeepSeek-v4-flash model and does not currently support the DeepSeek-v4-pro model.

Models supported by the Responses API The Responses API currently only supports the DeepSeek-v4-flash model and does not currently support the DeepSeek-v4-pro model. We will add support for the DeepSeek-v4-pro model in early August 2026. In order to meet everyone's needs for Codex, our API has added support for the Responses API format, and its base_url is. Through simple configuration, the DeepSeek model can be used in Codex. To connect the DeepSeek model to Codex, please refer to Connecting to Codex to call the DeepSeek model through Responses API # Please install OpenAI SDK first: `pip3 install OpenAI` from OpenAI import OpenAI client = OpenAI(api_key=" ", base_url=" response = client.responses.create( model="DeepSeek-v4-flash", instructions="You are a helpful assistant.", input="Hi, how are you?", ) print(response.output_text) Streaming Output With stream: true , responses will be returned as a semantically streaming sequence of SSE events. Each event is streamed with an event field indicating the event type and an incrementing sequence_number . response.incomplete / response.failed event ends with no data: [DONE] message. stream = client.responses.create( model="DeepSeek-v4-flash", instructions="You are a helpful assistant.", input="Hi, how are you?", stream=Tr