Overview
GPT-4o was released
about 1 year after
GPT-4.
GPT-4o
|
GPT-4
|
|
|---|---|---|
|
Provider
The entity that provides this model.
|
OpenAI
|
OpenAI
|
|
Input Context Window
The number of tokens supported by the input context window.
|
128K
tokens
|
8,192
tokens
|
|
Maximum Output Tokens
The number of tokens that can be generated by the model in a single request.
|
2,048
tokens
|
8,192
tokens
|
|
Release Date
When the model was first released.
|
2024-05-13
|
2023-03-14
|
|
Knowledge Cutoff
Limit on the knowledge base used by the model.
|
October 2023
|
September 2021
|
|
Open Source
|
|
|
|
API Providers
The providers that offer this model. (This is not an exhaustive list.)
|
|
|
Pricing
GPT-4o is
roughly 6.0x cheaper compared
to GPT-4 for input tokens and
roughly 4.0x cheaper
for output tokens.
|
GPT-4o
|
GPT-4
|
|
|---|---|---|
|
Input
Cost of input data provided to the model.
|
$5.00
per million tokens
|
$30.00
per million tokens
|
|
Output
Cost of output tokens generated by the model.
|
$15.00
per million tokens
|
$60.00
per million tokens
|
Benchmarks
Compare relevant benchmarks between GPT-4o
and GPT-4.
|
GPT-4o
|
GPT-4
|
|
|---|---|---|
|
MMLU
Evaluating LLM knowledge acquisition in zero-shot and few-shot settings.
|
88.7
(5-shot)
|
86.4
(5-shot)
|
|
MMMU
A wide ranging multi-discipline and multimodal benchmark.
|
69.1
|
34.9
|
|
HellaSwag
A challenging sentence completion benchmark.
|
Benchmark not available.
|
95.3
(10-shot)
|
|
GSM8K
Grade-school math problems benchmark.
|
Benchmark not available.
|
92.0
(5-shot)
|
|
HumanEval
A benchmark to measure functional correctness for synthesizing programs from docstrings.
|
Benchmark not available.
|
67.0
(0-shot)
|
|
MATH
Benchmark performance on Math problems ranging across 5 levels of difficulty and 7 sub-disciplines.
|
76.6
(0-shot)
|
Benchmark not available.
|
Llama 3.1 405B Instruct
Mistral 8x7B Instruct
Claude 3 Sonnet
Measure & Improve LLM
Product Performance.
Get Started
© 2023-2024 Context