Claude 4 Opus

Reasoning

Creator

Anthropic

Released

2025-05-22

Intelligence

39.0

Artificial Analysis Index

Coding

34.0

Artificial Analysis Index

In $/1M

$18.75

input tokens

Out $/1M

$75.00

output tokens

Blended $/1M

$32.81

3:1 blended

Speed

45

tokens / sec

Benchmark breakdown

Independent evaluation scores, normalized to 0–100.

GPQA Diamond
80
Humanity’s Last Exam
12
MMLU-Pro
87
SciCode
40
LiveCodeBench
64
MATH-500
98
AIME 2025
73
τ²-Bench (agentic)
73
Terminal-Bench Hard
31
IFBench
54
Via Artificial Analysis