Anthropic

Claude Opus 5

Strongest Claude Opus model for coding, agents, and professional work

anthropic/claude-opus-5

Context

1M

Output

128K

Released

2026-07-24

Knowledge cutoff

2026-05

Types

InputTextImagePDF
OutputText

Reasoning

Supported

Tool use

Supported

Structured output

—

Attachments

Supported

Open weights

—

Provider pricing

Cloudflare AI Gateway
InTextImagePDF
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
CrossModel
InTextImage
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
Eden AI
InTextImagePDF
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
Kilo Gateway
InTextImagePDF
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
LLM Gateway
InTextImagePDF
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
Merge Gateway
InTextImagePDF
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
NanoGPT
InTextImagePDF
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
NEAR AI Cloud
InTextImage
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
Ofox
InTextImagePDF
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
OpenRouter
InTextImagePDF
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
OrcaRouter
InTextImagePDF
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
Prime Inference
InTextImage
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
Tempr Gateway
InTextImagePDF
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
Vercel AI Gateway
InTextImagePDF
OutText
Price
$5.00 / $25.00
Context
1M
Status
Stable
AI/ML API
InTextImagePDF
OutText
Price
$6.50 / $32.50
Context
1M
Status
Stable

Benchmarks

  • SWE-Bench Verified

    96 · resolved

  • SWE-Bench Pro

    79.2 · resolve rate

  • SWE-Bench Multilingual

    89.5 · resolve rate

  • SWE-Bench Multimodal

    59.4 · resolve rate

  • DeepSWE

    68.8 · resolve rate

  • FrontierCode

    53.4 · mean@5

  • Frontier-Bench

    43.3 · mean reward

  • BrowseComp

    90.8 · accuracy

  • Humanity's Last Exam

    56.3 · accuracy

  • Humanity's Last Exam

    64.7 · accuracy

  • DeepSearchQA

    95 · F1

  • OSWorld

    70.6 · success rate

  • GDPval-AA

    1861 · Elo

  • AA-Briefcase

    1720 · Elo

  • AutomationBench

    26 · success rate

  • ARC-AGI-1

    97.5 · accuracy

  • ARC-AGI-2

    90.4 · accuracy

  • ARC-AGI-3

    30.2 · RHAE

  • HealthBench Professional

    59.8 · score

From the same lab