GPT-5

GPT-5 is a flagship model from OpenAI designed for coding, reasoning, and agentic tasks across domains. It is optimized for coding and agentic tasks with higher reasoning capabilities and medium speed.

SWE-Lancer (IC-Diamond subset)

100.0%

i
COLLIE

99.0%

i
Tau2 Telecom

96.7%

i
OpenAI-MRCR: 2 needle 128k

95.2%

i
AIME 2025

94.6%

i
HumanEval

93.4%

i
HMMT 2025

93.3%

i
MMLU

92.5%

i
BrowseComp Long Context 128k

90.0%

i
BrowseComp Long Context 256k

88.8%

i
Aider-Polyglot

88.0%

i
OpenAI-MRCR: 2 needle 256k

86.8%

i
VideoMME w sub.

86.7%

i
GPQA

85.7%

i
MATH

84.7%

i
VideoMMMU

84.6%

i
MMMU

84.2%

i
CharXiv-R

81.1%

i
Tau2 Retail

81.1%

i
MMMU-Pro

78.4%

i
Graphwalks BFS <128k

78.3%

i
SWE-Bench Verified

74.9%

i
Graphwalks parents <128k

73.3%

i
Multi-Challenge

69.6%

i
MultiChallenge (o3-mini grader)

69.6%

i
Scale MultiChallenge

69.6%

i
ERQA

65.7%

i
Internal API instruction following (hard)

64.0%

i
Tau2 Airline

62.6%

i
BrowseComp

54.9%

i
FrontierMath

26.3%

i
Humanity's Last Exam

24.8%

i
HealthBench Hard

1.6%

i
FActScore

1.0%

i
LongFact Objects

0.8%

i
LongFact Concepts

0.7%

i

Pricing, uptime, and speed via OpenRouter — updated Jul 17, 2026, 04:19 AM.

Provider	Status	Input	Output	Limits	Uptime	Speed	Notes
OpenAI	available	$1.25/Mtok cache $0.13/Mtok	$10.00/Mtok	400K tokens context 128K tokens max output	99% 5m 98%	2,430 ms p50 TTFT 65 tok/s p50	cache $0.01/web search
Azure	available	$1.38/Mtok cache $0.14/Mtok	$11.00/Mtok	400K tokens context 128K tokens max output	—	—	cache $0.01/web search