ChatGPT voice and image input (GPT-4V)

OpenAI · 25 September 2023

ChatGPT adds spoken conversations and the ability to upload and discuss images for Plus and Enterprise users, via GPT-4V.

First time GPT-4 could see images in a consumer product. Voice was a chained speech-to-text, LLM, text-to-speech pipeline (GPT-4o later replaced that with one end-to-end model).

Date

Monday, 25 September 2023

Lab

OpenAI

Kind

feature

Access

app only

The chained-pipeline contrast is drawn from Simon Willison's GPT-4o write-up.

Sources

1. hidekazu-konishi.com/entry/openai_gpt_model_release_timeline.html
2. en.wikipedia.org/wiki/GPT-4
3. en.wikipedia.org/wiki/ChatGPT
4. simonwillison.net/2024/May/13/gpt-4o/

This record was checked against its sources on 6 October 2026. How we check