OpenAI logo

GPT-5

OpenAI · Released Aug 7, 2025

ProprietaryreasoningcodingCompare this model

$0.625

Input price / MTok

#9 of 27

$5

Output price / MTok

#14 of 27

400K

Context window

#19 of 28

107

Tokens / sec

#12 of 21

74.9

SWE-bench Verified

#7 of 12

88.4

GPQA Diamond

#8 of 19

94.6

AIME 2025

#2 of 10

84.2

MMMU

#1 of 8

Summary

GPT-5 is a proprietary language model from OpenAI, released on Aug 7, 2025.

At $0.625 input and $5 output per million tokens, it ranks #9 of 27 for input price, placing it mid-range on input cost.

Its context window of 400K tokens ranks #19 of 28 for context depth in this dataset.

It targets reasoning, coding workloads.

Throughput is measured at 107 tokens per second (ranked #12 of 21 for throughput).

Original GPT-5 release; now superseded by 5.4/5.5 family.

Benchmarks

BenchmarkScoreMax%RankSource
SWE-bench Verified74.910074.9%#7/12OpenAI (self-reported in GPT-5 system ca...
GPQA Diamond88.410088.4%#8/19OpenAI (self-reported in GPT-5 system ca...
AIME 202594.610094.6%#2/10OpenAI (self-reported in GPT-5 system ca...
MMMU84.210084.2%#1/8OpenAI (self-reported in GPT-5 system ca...

Similar Models

Compare all 6 models

Models with the most shared benchmark coverage and closest scores to GPT-5.

GPT-5
OpenAIo3
AnthropicClaude Sonnet 4.6
OpenAIo4-mini
GoogleGemini 2.5 Pro
AnthropicClaude Haiku 4.5
SWE-bench Verified#7 of 12
Real-world GitHub issue resolution
GPQA Diamond#8 of 19
Graduate-level science reasoning
AIME 2025#2 of 10
Advanced math competition problems
MMMU#1 of 8
Multimodal understanding across 30 subjects

Compare GPT-5 side by side

Pre-filled with 5 similar models

Open comparison

Speed source: https://artificialanalysis.ai/models/gpt-5