Skip to content
llmwaves

GPT-4

OpenAI · released May 28, 2023

ProprietaryTool useStructured output

OpenAI's flagship model, GPT-4 is a large-scale multimodal language model capable of solving difficult problems with greater accuracy than previous models due to its broader general knowledge and advanced reasoning...

Specification

Context window
8K
Max output
4K
Knowledge cutoff
Sep 30, 2021
Parameters
Undisclosed
Licence
Proprietary
Serving providers
2
Moderated
No
Uptime
100.0%

Intelligence

6.7

20th percentile

Coding

13.1

Coding Index

Agentic

—

Agentic Index

Output speed

48 t/s

Median across providers

Latency

803ms

Time to first token

Cost per task

$3.00

Estimated

Benchmarks

Where the score comes from

The Intelligence Index is a composite. These are the underlying evaluations this model was actually measured on.

Evaluation scores

Percentage correct · higher is better

  • MMLU-Pro
    56.2%
  • GPQA Diamond
    34.9%
  • IFBench
    33.2%

An evaluation missing from this list was not run for this model — it is not a zero.

View as table
EvaluationScore
MMLU-Pro56.2%
GPQA Diamond34.9%
IFBench33.2%

Against its peers

Intelligence Index · this model highlighted, nearest peers in grey

Peers are the models sitting closest on the Intelligence Index — the set you would realistically choose between.

View as table
ModelIntelligence
Nemotron 3 Nano 30B A3B6.8
Mistral Large 24076.8
GPT-46.7
GLM 4.5V6.7
Gemini 2.5 Flash Lite6.7
Qwen2.5 Coder 32B Instruct6.7
GPT-4o-mini6.7
GPT-4o-mini (2024-07-18)6.7

Percentile among all indexed models

Intelligence20th
Coding13th

Pricing

What it costs to run

List prices per million tokens, plus what one representative task works out to.

List price

Input / 1M tokens
$30
Output / 1M tokens
$60
Cached input / 1M
Not offered
Blended 3:1
$37.5

One task, estimated

$3.00

Input tokens
50,000
Output tokens
25,000
Profile
Standard

Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).

Serving providers

Speed and latency figures are medians across these providers, so a widely-served model reports a blend rather than any single endpoint.