Latest / Frontier models and safety

Model releaseFrontierUSConfirmed

OpenAI releases GPT-4, a model that takes text and image input

See how this story connects

OpenAI released GPT-4 on 14 March 2023 as a large multimodal model that accepts text and images and returns text. OpenAI said it scored around the top 10% of test takers on a simulated bar exam, against the bottom 10% for GPT-3.5. Text input went to ChatGPT Plus subscribers with a usage cap and to API users from a waitlist, while image input stayed a research preview with Be My Eyes.

Why it matters

GPT-4 became the reference point for the most capable models, and the pause letter eight days later used it as the threshold for its proposed halt.

When
14 Mar 2023
ID
WTA-N-2023-0021
Who
OpenAI, Be My Eyes

Summary by wheresthe.ai, from OpenAI. Checked against the source on 9 Oct 2026.

Source copy: download the page text as saved on 18 Oct 2026 (PDF). Also held by the Internet Archive.

Next stop on this Storyline: What happened nextGoogle launches Gemini 1.0 in Ultra, Pro and Nano sizes

Cite this story

WTA-N-2023-0021: “OpenAI releases GPT-4, a model that takes text and image input.” wheresthe.ai, 14 Mar 2023. https://wheresthe.ai/d/openai-gpt-4-release-2023-03/ (source: OpenAI: GPT-4).

Spotted an error? Email [email protected].