readable AI benchmarks

Anthropic

Claude Opus 5 (Adaptive Reasoning, Low Effort)

reasoning · closed weights · Jul 24, 2026

Quality Score68.1
Value Score41.0
Reliability64.3
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
$0.425

Capability

Intelligence
52.5
Coding
66.9
Agentic
42.1
Omniscience
28.6
Correct
56.0%
Blended price
$3.85

Answer outcomes

Correct 56.0%Incorrect 27.4%Abstained 16.7%

Artificial Analysis benchmarks

GDPval-AA v2
48%
τ³-Banking
30%
Terminal-Bench v2.1
76%
SciCode
48%
Humanity’s Last Exam
43%
GPQA Diamond
89%
CritPt
23%
AA-Omniscience
64%
AA-LCR
77%

Similar models