Longterm Wiki

GPT-4o

OpenAI

GPT-4o ("o" for "omni") was released May 13, 2024 as OpenAI's flagship multimodal model. It natively processes text, image, and audio inputs and outputs. Matched GPT-4 Turbo intelligence at 50% lower price (\$2.50/\$10 per MTok) and 2x faster speed. Scored 88.7% on MMLU. Available free to all ChatGPT users with rate limits.

Source →
Developer
OpenAI
Released
2024-05-13
Context Window
128K tokens

Pricing

TypePrice per MTok
Input$2.5
Output$10

Benchmarks2

BenchmarkScore
MMLU88.7%
HumanEval90.2%

GPT Family6

ModelTierReleasedInput $/MTok
GPT-4.1flagship2025-04-14$2
GPT-4.1 minimini2025-04-14$0.4
GPT-4.1 nanonano2025-04-14$0.1
GPT-4o minimini2024-07-18$0.15
GPT-4 Turboflagship2023-11-06$10
GPT-4flagship2023-03-14$30

Other OpenAI Models6

o32025-04-16
o-series (flagship)
o4-mini2025-04-16
o-series (mini)
o3-mini2025-01-31
o-series (mini)
o12024-12-05
o-series (flagship)
o1-preview2024-09-12
o-series (preview)
o1-mini2024-09-12
o-series (mini)

Details

Model FamilyGPT
Tierflagship
Generation4o
Release Date2024-05-13
Context Window128K tokens

Capabilities4

tool-usevisionaudio-inputaudio-output

Sources1

Hello GPT-4o

Tags

gptgpt-4omultimodal