What is the fundamental difference between ChatOpenAI("gpt-3.5-turbo") and OpenAI("gpt-3.5-turbo")?

[samurai00]

What is the fundamental difference between ChatOpenAI("gpt-3.5-turbo") and OpenAI("gpt-3.5-turbo")?
What are the differences in interaction with the OpenAI API?
Which one is more recommended to use in ConversationalRetrievalQAChain?

I am confused about the selection and use of these two models.
I want to custom prompts for ConversationalRetrievalQAChain and summarization chain, but the prompt formats used by these two models seem to be different.

Example prompt selector in the project :

const qa_template = `Use the following portion of a long document to see if any of the text is relevant to answer the question. 
Return any relevant text verbatim.
{context}
Question: {question}
Relevant text, if any:`;
export const DEFAULT_COMBINE_QA_PROMPT =
  /*#__PURE__*/
  PromptTemplate.fromTemplate(qa_template);

const system_template = `Use the following portion of a long document to see if any of the text is relevant to answer the question. 
Return any relevant text verbatim.
----------------
{context}`;
const messages = [
  /*#__PURE__*/ SystemMessagePromptTemplate.fromTemplate(system_template),
  /*#__PURE__*/ HumanMessagePromptTemplate.fromTemplate("{question}"),
];
const CHAT_QA_PROMPT =
  /*#__PURE__*/ ChatPromptTemplate.fromPromptMessages(messages);

export const COMBINE_QA_PROMPT_SELECTOR =
  /*#__PURE__*/ new ConditionalPromptSelector(DEFAULT_COMBINE_QA_PROMPT, [
    [isChatModel, CHAT_QA_PROMPT],
  ]);