Anthropic

Claude Opus 4.7

Name: Anthropic Claude Opus 4.7
Author: Anthropic

Compare

2026-04-16

Compare

Claude Opus 4.7 is Anthropic's flagship frontier model, released April 16, 2026 as a successor to Opus 4.6. It is Anthropic's strongest coding and agentic model to date, purpose-built for long-horizon autonomous work with minimal supervision. Opus 4.7 plans deliberately, writes verification steps (tests, runs, fixes) before declaring tasks complete, and uses file-system memory to carry context across sessions. New controls include task budgets and an xhigh effort level between high and max. At launch it topped SWE-bench Verified at 87.6% and SWE-bench Pro at 64.3%, ahead of GPT-5.4 and Gemini 3.1 Pro, with 69.4% on Terminal-bench 2.0 and ~77% on ARC-AGI-2. Vision capabilities tripled, enabling 98.5% computer-use visual acuity. It features a 1M-token context window and 128K output capacity.

VisionReasoningWeb SearchFile|Proprietary Model

Knowledge Cutoff

2026-01-01

The date this AI finished learning. It may not know about things that happened after this date.

Input → Output Format

The types of content this AI can receive, and what it can produce in return.

Context Memory

1MIN128KOUT

The maximum amount of text the AI can read and process in a single request. A larger number means it can handle longer documents or conversations.

Cost/1M Words

$5IN$25OUT

The cost of using this AI directly in your own application. Shown in USD per 1 million units of text (tokens).

Calculate Cost

Source:Official Docs

AI Performance Evaluation

Arena Overall Score

1503

±7

As of 2026-05-01

Overall Rank

🥇 No.1

7,615 Votes

Arena by Ability

Hard Prompts

1530±9🥉 No.3

Expert Knowledge

1544±21🥇 No.1

Instruction Following

1504±12🥈 No.2

Conversation Memory

1526±18🥇 No.1

Creative

1494±17🥇 No.1

Coding

1573±14🥇 No.1

Math

1501±27No.6

Arena by Occupation

Creative Writing

1501±14🥇 No.1

Social Sciences

1520±16🥇 No.1

Media

1479±16🥉 No.3

Business

1494±15No.5

Healthcare

1511±25No.7

Legal

1500±24No.8

Software

1557±11🥇 No.1

Mathematics

1509±27No.8

Source:Arena Intelligence

Overall

AA Intelligence Index

57%↑18%

LiveBench

68%↑8%

Reasoning & Math

GPQA Diamond

91%↑9%

HLE

40%↑22%

LB Reasoning

75%↑6%

LB Math

76%↑2%

LB Data

76%↑22%

Coding

AA Coding Index

53%↑16%

LB Coding

76%↑4%

LB Agentic

60%↑15%

TAU2

89%↑8%

TerminalBench

52%↑17%

SciCode

55%↑13%

Language & Instructions

IFBench

59%↓5%

AA-LCR

70%↑8%

Hallucination (HHEM)

12%↑2%

Factual (HHEM)

88%↓2%

LB Language

75%↑2%

LB IF

46%↓5%

Output Speed

Standard Mode

49tok/s↓28

First Output 1.64s

Reasoning Mode

55tok/s↓31

First Output 23.21s

Source:Artificial Analysis LiveBench Vectara HHEM

Anthropic