Artificial intelligence · 2023

GPT-4

A multimodal model scores around the top 10% on a simulated bar exam.

Released March 14, 2023, GPT-4 accepted images as well as text and scored about 86% on MMLU. It jumped from GPT-3.5's bottom 10% to roughly the top 10% of a simulated Uniform Bar Exam. It became the capability bar every lab chased for a year.

See it in the living tree Wikipedia

Arcs above join it to what it stood on, arcs below to what it led to. The bar places this stretch on the tree's whole timeline.

What it built on 2

ChatGPT 2022
Neural scaling laws 2020

What it led to 4

AutoGPT 2023
Gemini 1.0 2023
Vision-language-action models 2023
GPT-4o 2024

Watch

GPT-4 Developer LivestreamOpenAI · YouTube

Related measurements

The dashed line marks when it first worked.

Best AI score on GPQA Diamond Latest: 95.8 % correct (2026) Source: Epoch AI, Capabilities & benchmarking (benchmark_data.zip, gpqa_diamond.csv); Anthropic: Claude 3 public launch (March 4, 2024)

Training compute of notable AI models Source: Epoch AI, Notable AI Models (CC BY 4.0)