Skip to main content
Anthropic logo

Anthropic

AI research and products that put safety at the frontier

Inference specialist🇺🇸 USinferencesafetyenterprise

Last reviewed Mar 14, 2026

Anthropic is an AI safety company that builds reliable, interpretable, and steerable AI systems. Operating as a public benefit corporation, Anthropic offers the Claude family of large language models via API, emphasizing responsible AI development through Constitutional AI.

LLM Models
22
From / 1M input
$0.50

LLM API Pricing

Pay-per-token pricing. Prices shown per 1M tokens.

Prices last updated: July 27, 2026

Pricing
ModelInput/1MOutput/1M
$1.00$5.00
$2.00$10.00
$3.00$15.00
$3.00$15.00
$5.00$25.00
$5.00$25.00
$5.00$25.00
$5.00$25.00
$5.00$25.00
$10.00$50.00

Pros & Cons

Advantages

  • Clean API design with SDKs for Python, TypeScript, Go, Java, Ruby, PHP, and C#
  • 200K token context window across all plans
  • Cost savings via prompt caching and the Batch API
  • Comprehensive documentation with quickstarts, courses, and a cookbook
  • Safety focus with Constitutional AI approach
  • Comprehensive enterprise features and integrations
  • Agent capabilities via Claude Code, Cowork, and Managed Agents

Limitations

  • No image or video generation capabilities
  • Higher cost for top-tier Fable and Opus models
  • Closed source proprietary models
  • Newer enterprise features may require learning curve

Key Features

Claude Model Family

Access to the Claude model family including Fable 5, Opus 5, Sonnet 5, and Haiku 4.5, plus the Mythos line

Claude Code

Agentic coding assistant in your terminal for enhanced development workflows

Claude Cowork

Collaborative workspace features for team projects and shared workflows

Claude Design

Design-focused product surface included with paid Claude plans

Claude Science

Customizable app for researchers that integrates common research tools and packages, produces auditable artifacts, and provides flexible access to computing resources

Claude Security

Security features and tools for enterprise security teams

Claude Managed Agents

Fully managed agent infrastructure with stateful sessions and persistent event history

Large Context Windows

200K tokens standard with extended context options for large document analysis

Advanced Tool Use

Web search, code execution, computer use, bash tool, memory, and skills capabilities

Structured Outputs

Schema-constrained responses for reliable machine-readable output

Enterprise Integrations

Claude for Chrome, Slack, Microsoft 365, and Microsoft Outlook, plus remote MCP connectors

Prompt Caching

Up to 90% cost savings on repeated content with cache durations

Vision Support

Process images and PDF documents natively

Batch API

50% cost reduction for asynchronous processing

Pricing Options

OptionDetails
Pay-per-token APIPer million token pricing for Claude models with competitive rates
Consumer PlansFree, Pro, and Max subscription tiers for individual use, billed monthly or annually
Team PlansPer-seat pricing with monthly or annual billing, central billing, and SSO
Enterprise PlansSelf-serve and sales-assisted per-seat pricing with advanced admin, security, and compliance features, also available via AWS Marketplace
Prompt Caching90% savings on cached content with 5-minute and 1-hour options
Batch API50% discount on all tokens for async processing

Availability & Support

Regions

150+ countries including US, Canada, UK, EU, Australia, Japan. Available via direct API, Amazon Bedrock, Google Cloud Vertex AI, and Microsoft Foundry

Support

Documentation, Discord community, email support, Help Center, status page, and enterprise support options

Getting Started

  1. 1

    Create Console account

    Sign up at platform.claude.com

  2. 2

    Generate API key

    Create an API key from Account Settings

  3. 3

    Install SDK

    pip install anthropic (Python) or npm install @anthropic-ai/sdk (TypeScript)

  4. 4

    Make first API call

    Call the Messages API endpoint with your API key

Compare Providers

Find the best prices for the same GPUs and models from other providers