What is the difference between ChatGPT dictation and voice mode?

ChatGPT dictation and voice mode both use your microphone, but they do different jobs. Dictation records a prompt, turns it into text, and puts that text in the box for you to edit and send. Voice mode is a spoken conversation: you talk, and the app talks back. If you want to read the prompt before the model sees it, use dictation. If you want to talk through a problem and hear the answer, use voice mode.

Dictation gives you text to review, and voice mode gives you a conversation

OpenAI's ChatGPT Voice help article draws the line in one sentence: "Use ChatGPT Dictation when you want to record a prompt, review and edit its transcription, and then send it as text." ¹ The article's point is the review step. The words stay in the box until you choose to send them.

OpenAI's developer walkthrough says the same thing in a different way. Voice dictation is for turning speech into prompt text before you send it, and it is separate from ChatGPT Voice, which is a live spoken conversation ². Two features with two buttons is the distinction that matters when something goes wrong, because a fault in one does not mean the other is broken.

Compare the two on the points that change your choice

The last two rows are the ones to weigh. Dictation fits a prompt you want to be exact, because you can fix a misheard name or a wrong number before the model reads it. Voice mode fits a conversation where a fast reply matters more than an exact first draft.

The audio row needs care. For dictation we can cite OpenAI's own answer: the recording goes to the models for transcription, and it stays linked to your chat history. OpenAI's FAQ says that the audio "will be retained for as long as the chat is part of your chat history," and that deleting the chat removes the clip within 30 days, unless OpenAI must keep it for security or legal reasons ⁴. We did not find a matching page for voice mode in the sources we read, so we do not state its retention.

How to decide in one question

Ask whether you want to read the prompt before the model sees it. If yes, use dictation. The prompt appears as text, and you press send when you are satisfied. If no, and you want to hear the reply and keep talking, use voice mode.

A few situations are clear cut:

• A long prompt with a reference you must get right, such as a product name or an email address: dictation, then read the text.
• A quick question while your hands are busy: voice mode, since the reply comes back as speech.
• A draft you will later paste into a document: dictation, so the text is ready to copy.

Claude has the same split

Claude has the same distinction. Anthropic's Claude help material describes dictation as converting speech to text so you can type prompts by speaking, while voice mode is a two-way spoken conversation ⁵. In Claude Code, dictation is available but voice mode is not ⁶, so the terminal tool gives you only the text route.

For the steps in each app, see our guide to dictation in ChatGPT and Claude. If the button you need is not working, the steps are in why ChatGPT dictation is not working and why Claude dictation is not working.

What the answer does not cover

Both features send your voice to OpenAI, so privacy is a fair question for either one. For a general view of how to think about it, read our guide to whether voice to text is safe. The apps and browsers overview lists the other guides in this group, including the one on which speech model ChatGPT uses.