readable AI benchmarks

Anthropic

Claude Opus 5 (Adaptive Reasoning, High Effort)

reasoning · closed weights · Jul 24, 2026

Quality Score76.2
Value Score46.7
Reliability66.9
Cache Discount90%

Model specification

Reasoning
reasoning
Input modalities
text, image
Output modalities
text
Context window
1M tokens
Weights
closed weights

Token prices USD per 1M tokens

Input
$5.00
Output
$25.00
Cached Input
$0.500
Cached Output
$6.25
Cost per task
$1.23

Capability

Intelligence
61.5
Coding
76.5
Agentic
56.1
Omniscience
33.7
Correct
58.9%
Blended price
$3.85

Answer outcomes

Correct 58.9%Incorrect 25.2%Abstained 16.0%

Artificial Analysis benchmarks

GDPval-AA v2
62%
τ³-Banking
45%
Terminal-Bench v2.1
88%
SciCode
54%
Humanity’s Last Exam
53%
GPQA Diamond
94%
CritPt
28%
AA-Omniscience
67%
AA-LCR
76%

Similar models