readable AI benchmarks

Anthropic

Claude Opus 5 (Adaptive Reasoning, Medium Effort)

reasoning · closed weights · Jul 24, 2026

Quality Score73.1
Value Score44.7
Reliability65.5
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
$0.724

Capability

Intelligence
58.6
Coding
74.3
Agentic
50.4
Omniscience
31.0
Correct
57.1%
Blended price
$3.85

Answer outcomes

Correct 57.1%Incorrect 26.0%Abstained 16.9%

Artificial Analysis benchmarks

GDPval-AA v2
56%
τ³-Banking
39%
Terminal-Bench v2.1
86%
SciCode
51%
Humanity’s Last Exam
51%
GPQA Diamond
92%
CritPt
27%
AA-Omniscience
66%
AA-LCR
79%

Similar models