ModelBeat
Models/Claude Opus 4.8
All models
A

Anthropic: Claude Opus 4.8

anthropic/claude-opus-4.8

Claude Opus 4.8 is Anthropic's most capable generally available model in the Opus family.

Context
1M
Input
$5 / 1M
Output
$25 / 1M
Released
May 28, 2026
Modalities
textimagefiletext

Benchmarks

Scores on standardized evaluations. Higher is better — and rank shows where Claude Opus 4.8 lands among all models tracked on Model Beat.

89.0
Intelligence Index
Epoch AI
89th percentile of tracked models
96.0
Coding Index
Epoch AI
96th percentile of tracked models
94.0
Agentic Index
Epoch AI
94th percentile of tracked models

Reasoning

6 evals
ARC-AGI92.5%

Abstract visual reasoning

ARC-AGI-272.1%

Harder abstract reasoning

GPQA Diamond91.0%

Graduate-level scientific reasoning

SimpleBench64.8%

Common-sense trick questions

SimpleQA Verified39.5%

Factual accuracy & hallucination

WeirdML82.9%

Novel ML problem-solving

Coding

1 evals
WebDev Arena1552

Human-rated web development

Math

3 evals
AIME 2024/202598.3%

Olympiad-qualifier math

FrontierMath47.2%

Research-level math problems

FrontierMath Tier 431.3%

Hardest research math

Agentic & Tools

1 evals
APEX42.5%

Multi-step agentic tasks

In the news

1

Model & benchmark data from Epoch AI (CC BY); pricing, specs & descriptions from OpenRouter.