Claude Opus 5

Claude Opus 5 is Anthropic's flagship model for advanced coding, long-running agentic tasks, research, and complex knowledge work.

Reasoning
Vision
Native PDF input
Tool Calling
Structured Output

Added Jul 24, 2026

Pricing

Auto routing · per 1M tokens

Input: $5.00
Output: $25.00
Cache read: $0.50

Compare provider prices

Specifications

Context window: 1M
Max output: 128K
Avg output (7d): 577 tokens; Longer than 51% of models

Benchmarks

Artificial Analysis

LMArena

Design Arena

Sourced from Artificial Analysis.

Intelligence Index

48.1

Better than 97% of models compared

Coding Index

76.5

Better than 95% of models compared

Agentic Index

52.3

Better than 94% of models compared

Agentic work

AutomationBench-AA

Workflow automation with guardrail penalties

53.6%

Better than 71% of models compared

AA-Briefcase

Agentic knowledge work (Elo)

1562 Elo

Better than 91% of models compared

GDPval-AA v2

Economically valuable tasks (Elo)

1596 Elo

Better than 89% of models compared

Document reasoning

GDP.pdf

Professional PDF reasoning: all-pass rate

19.6%

Better than 69% of models compared

AA-LCR v1.1

Long context reasoning with updated grading

79.0%

Better than 82% of models compared

MLCR-AA

Medical long-context reasoning

59.4%

Better than 95% of models compared

Reasoning

HLE

Humanity's Last Exam

52.8%

Better than 97% of models compared

CritPt

Research-level physics reasoning

28.3%

Coding

Terminal-Bench v4.0

Practical coding and terminal tasks

46.0%

Better than 90% of models compared

SciCode

Python programming for scientific computing

55.4%

Better than 79% of models compared

Knowledge

AA-Omniscience Accuracy

Proportion of correctly answered questions

58.9%

AA-Omniscience Hallucination Rate

Rate of incorrect answers among non-correct responses

61.2%

Legacy benchmarks

GPQA Diamond (legacy)

Graduate-level scientific reasoning

93.7%

Better than 98% of models compared

AA-LCR (unversioned / legacy)

Long context reasoning evaluation

79.0%

Better than 82% of models compared

GDPval-AA (unversioned / legacy)

Economically valuable tasks

54.7%

Last updated Oct 14, 2026