Bài kiểm tra có liên kết hướng dẫn · Trung bình Cấp độ
API phản hồi OpenAI so với bài kiểm tra hoàn thành trò chuyện
So sánh hướng dẫn điểm cuối OpenAI hiện tại, hình dạng yêu cầu và phản hồi, tùy chọn trạng thái, công cụ tích hợp, khả năng tương thích mô hình và công việc di chuyển.
Câu hỏi 1 của 8
As of the cited OpenAI documentation, how can Responses preserve conversation context across turns?
Xem trước mọi câu hỏi8 câu hỏi
Câu trả lời và giải thích xuất hiện khi bạn làm bài kiểm tra.
- As of the cited OpenAI documentation, how can Responses preserve conversation context across turns?
- Through options such as previous_response_id or the Conversations API
- Only by switching to Chat Completions
- Automatically across every API call and user
- By relying on the model to remember earlier requests
- What built-in capability does the Responses API offer that previously required custom implementation in Chat Completions?
- Streaming text output
- Built-in tools such as web search and file search executed server-side
- Setting a system prompt
- Assigning roles to messages
- How does the output shape of the Responses API differ from Chat Completions?
- It returns a single flat message just like Chat Completions
- It only ever returns tool call results, never text
- It can return multiple output items such as text, tool calls, and reasoning traces
- It removes the concept of roles entirely
- What does OpenAI currently say about Chat Completions?
- It is already shut down
- It is required for all built-in tools
- It is the only API that supports text generation
- It remains supported, while Responses is recommended for new projects
- What should a developer check before assuming an API feature will work with a particular model?
- The model-specific endpoint and tool support in current docs
- Only the marketing name
- That every model supports all endpoints identically
- Whether another provider supports the feature
- In custom function calling under Chat Completions, who executes the requested function?
- OpenAI's servers execute it automatically
- The developer's own application code executes it and sends back the result
- The function runs inside the model's weights
- No execution is needed; the model fabricates a plausible result
- What does previous_response_id allow a developer to skip?
- Setting a system prompt
- Choosing a model name
- Resending the full conversation history on each request
- Enabling streaming
- Which task is part of a careful Chat Completions-to-Responses migration?
- Changing only the URL and assuming all fields match
- Removing application-level safety checks
- Assuming Responses retains no data
- Adapting input/output parsing and testing tools, streaming, errors, and storage settings