Question

My script stopped working after OpenAI switched to the new Responses API How do I migrate from ChatCompletion

Answer 1

The responses API is quite different from old.completions
The old version had
from openai import OpenAI

client = OpenAI()

response = client.chat.completions.create(

    model="gpt-4o",

    messages=[

        {"role": "system", "content": "You are a helpful assistant."},

        {"role": "user", "content": "Hello!"}

    ]

)

print(response.choices[0].message.content)
The new version is
from openai import OpenAI

client = OpenAI()

response = client.responses.create(

    model="gpt-4o",

    instructions="You are a helpful assistant.",  # replaces system message

    input="Hello!"  # replaces messages array for simple cases

)

print(response.output_text)

So the basic difference is
1. System inputs : Instructions parameter , this is not a message array .
2. Message -input : for single turn calls , it accepts a single plain array , for multiturn , it accepts arrays similar to before .
3. Response shape changed

• Old: response.choices[0].message.content

• New: response.output_text (for simple text) or response.output[0].content[0].text .