Claude Opus 4.8
Claude Opus 4.8 (Adaptive Reasoning, Max Effort) is an AI model from Anthropic.
Cite
Notes
Only stored in your browser.
Context
1M
Input
$5
/1M tok
Output
$25
/1M tok
Cache read
$0.50
0.10x input
Cache write
$6.25
1.25x input
Price & speed via Artificial Analysis · ranked across tracked models · caching via OpenRouter
Reported on 42 evals across 23 domains - best 94.4% on τ²-bench (Tau²-bench) · #27 of 321
Reported eval scores
42IFBench62.2
LiveBench - Math84.3
SAGE54.8
LiveBench80.1
Vals Index55.1
SciCode53.5
MortgageTax69.9
TaxEval v275.6
CorpFin v266.7
OSWorld-Verified83.4
Head-to-heads
29benchmark comparisons against other models
vs GPT-5.541 shared evalsvs Claude Fable 540 shared evalsvs Gemini 3.1 Pro Preview38 shared evalsvs Claude Opus 4.730 shared evalsvs Claude Sonnet 4.629 shared evalsvs Claude Opus 4.628 shared evalsvs GPT-5.428 shared evalsvs Kimi K2.528 shared evalsvs GLM 5.127 shared evalsvs GPT-5.226 shared evalsvs GPT-5.4 Mini26 shared evalsvs Claude Sonnet 4.525 shared evalsvs Qwen3.6 Plus25 shared evalsvs GPT-5 Mini24 shared evalsvs Claude Opus 4.524 shared evalsvs Grok 4.1 Fast24 shared evalsvs GPT-5 Nano22 shared evalsvs Claude 4 Sonnet22 shared evalsvs DeepSeek V3.222 shared evalsvs GLM 4.721 shared evalsvs Gemini 2.5 Pro20 shared evalsvs GPT-6 Astra19 shared evalsvs gpt-oss-120b18 shared evalsvs GPT-517 shared evalsvs GPT-4o14 shared evalsvs GPT-4o-mini12 shared evalsvs GPT-4.1 Nano12 shared evalsvs GPT-4.1 Mini11 shared evalsvs GPT-4.111 shared evals