GPT-4.1

GPT-4.1 is OpenAI's latest and most advanced flagship model, significantly improving upon GPT-4 Turbo in performance across benchmarks, speed, and cost-effectiveness.

MMLU

90.2%

i
CharXiv-D

87.9%

i
IFEval

87.4%

i
MMMLU

87.3%

i
MMMU

74.8%

i
MathVista

72.2%

i
Video-MME (long, no subtitles)

72.0%

i
Multi-IF

70.8%

i
TAU-bench Retail

68.0%

i
GPQA

66.3%

i
COLLIE

65.8%

i
ComplexFuncBench

65.5%

i
Graphwalks BFS <128k

61.7%

i
Graphwalks parents <128k

58.0%

i
OpenAI-MRCR: 2 needle 128k

57.2%

i
CharXiv-R

56.7%

i
SWE-Bench Verified

54.6%

i
Aider-Polyglot Edit

52.9%

i
Aider-Polyglot

51.6%

i
TAU-bench Airline

49.4%

i
Internal API instruction following (hard)

49.1%

i
AIME 2024

48.1%

i
AIME 2025

46.4%

i
OpenAI-MRCR: 2 needle 1M

46.3%

i
MultiChallenge (o3-mini grader)

46.2%

i
Multi-Challenge

38.3%

i
HMMT 2025

28.9%

i
Graphwalks parents >128k

25.0%

i
Graphwalks BFS >128k

19.0%

i
Humanity's Last Exam

5.4%

i

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

Provider	Status	Input	Output	Limits	Uptime	Speed	Notes
OpenAI	available	$2.00/Mtok cache $0.50/Mtok	$8.00/Mtok	1.0M tokens context 33K tokens max output	99.3% 5m 98%	549 ms p50 TTFT 20 tok/s p50	$0.01/web search
Azure	available	$2.20/Mtok cache $0.55/Mtok	$8.80/Mtok	1.0M tokens context 33K tokens max output	—	—	cache $0.01/web search