0

Claude Opus 4.8

Claude Opus 4.8 (Adaptive Reasoning, Max Effort) is an AI model from Anthropic.

Publisher
Anthropic
Family
Claude
Context
1M
Input
Output
Released
28 May 2026

Cite

Notes

Only stored in your browser.

Attribution

Benchmark scores
AALiveBenchVals AI
Attribution policy →
Context
1M
Input
$6.25
/1M tok
Output
$25
/1M tok
Blended
$10.94
#399/ 414/1M · 3:1
Speed
60
#160/ 414tok/s
Cache read
$0.50
0.08x input
Cache write
$6.25
1.00x input

Price & speed via Artificial Analysis · ranked across tracked models · caching via OpenRouter

Reported on 24 evals across 10 domains - best 94.4% on τ²-bench (Tau²-bench) · #15 of 320

Reported eval scores

24
LiveBench - Reasoning89.7
Terminal-Bench (Hard)58.3
LiveBench - Data Analysis78.3
LiveBench - Language81.4
LiveBench - Coding79.3
IFBench62.2
LiveBench - Math84.3
LiveBench - Instruction Following67.4
Vals Index70.2
Humanity's Last Exam (HLE)45.7
LiveBench80.1
Vals Multimodal Index70.7
Vibe Code Bench v1.182.7
ProofBench69
GPQA Diamond92
SciCode53.5
MortgageTax69.9
Finance Agent (v2)53.9
TaxEval v275.6
CorpFin v266.7
SAGE54.8
τ²-bench (Tau²-bench)94.4
MedScribe85.8
MedCode53.2