Claude Opus 5

Adaptive Reasoning, Medium Effort

Creator

Anthropic

Released

2026-07-24

Intelligence

56.3

Artificial Analysis Index

Coding

74.3

Artificial Analysis Index

In $/1M

$5.00

input tokens

Out $/1M

$25.00

output tokens

Blended $/1M

$10.00

3:1 blended

Speed

54

tokens / sec

Benchmark breakdown

Independent evaluation scores, normalized to 0–100.

GPQA Diamond
92
Humanity’s Last Exam
49
MMLU-Pro
SciCode
51
LiveCodeBench
MATH-500
AIME 2025
τ²-Bench (agentic)
Terminal-Bench Hard
IFBench
Via Artificial Analysis
© 2026 NYSGPT2525 LLC