What is Claude
Claude has built a strong reputation around careful writing, long-context work, structured reasoning, and a calm, dependable user experience. Teams often shortlist Claude when they care about quality of output, document-heavy workflows, and a product that feels less noisy than many AI alternatives.
Claude comes from Anthropic, which was founded in 2021 and quickly built its reputation around AI safety, clear reasoning, and dependable writing quality. Since launching Claude publicly, Anthropic has expanded from a respected lab into one of the most important commercial AI vendors for businesses that care about trust, long-context work, and careful output.
Core offerings
- Claude chat for writing, drafting, research, analysis, and internal knowledge work.
- API access for product teams and developers.
- A model family that spans higher-end reasoning and premium quality down to faster, cheaper options.
- Team and enterprise plans for workspace management and broader rollout.
Pricing
Claude Free is US$0. Pro costs US$20 monthly or US$17 per month effective when US$200 is paid annually. Max remains available in 5x and 20x usage tiers at US$100 and US$200 monthly. Team now supports 2–150 users: Standard costs US$25 per member monthly or US$20 on annual billing, while Premium costs US$125 monthly or US$100 annually. Enterprise self-serve is US$20 per seat per month plus usage at API rates; sales-assisted Enterprise uses custom pricing.
For developers, Claude Fable 5.1 is US$10 per million input tokens and US$50 per million output tokens, with US$0.25 cache reads and US$12.50 five-minute cache writes. Claude Opus 5.5 is US$4 input and US$20 output with US$0.20 cache reads and US$5 five-minute cache writes; Claude Sonnet 5.5 is US$2 and US$10 with US$0.10 cache reads and US$2.50 five-minute cache writes. Claude Haiku 5.5 uses tiered rates: up to 100,000 input tokens it is US$0.10 input and US$0.50 output, while longer prompts are US$0.50 input and US$2.50 output per million tokens. Anthropic separately publishes optional tool charges, including managed agents at US$0.08 per session-hour, web search at US$10 per 1,000 searches, and additional code-execution usage at US$0.05 per hour.
Model footprint
Anthropic’s public pricing covers Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5. Claude Cowork is now presented simply as Claude as Anthropic rolls the unified product experience across paid plans.
Why select Claude
Claude is a strong option for organisations that care about writing quality, reasoning quality, long-context document analysis, and a cleaner product surface. It is often especially attractive for policy, legal, research, operations, and internal knowledge workflows.
Official sources: Claude plans, Claude API pricing, and Claude models.
Current models
Anthropic publishes current model prices and limits, while training-data specifics remain undisclosed.
| Model | Context / output | Knowledge / training | Pricing / token use | Speed / notes |
|---|---|---|---|---|
| Claude Fable 5.1 | See current model card | Training data: not publicly disclosed | US$10 input / US$50 output per 1M tokens; US$12.50 cache write / US$0.25 cache read | Anthropic's newest model for long-running agents |
| Claude Opus 5.5 | 1M context | Training data: not publicly disclosed | US$4 input / US$20 output per 1M tokens; US$5 cache write / US$0.20 cache read | Current premium tier for high-complexity reasoning and agentic work |
| Claude Sonnet 5.5 | 1M context | Training data: not publicly disclosed | US$2 input / US$10 output per 1M tokens; US$2.50 five-minute cache write / US$4 one-hour cache write / US$0.10 cache read | Current balanced tier for coding, agents, and general professional work |
| Claude Haiku 5.5 | Tiered pricing above 100K input tokens | Training data: not publicly disclosed | Up to 100K input: US$0.10 input / US$0.50 output per 1M tokens; above 100K: US$0.50 input / US$2.50 output; cache reads are US$0.01 or US$0.05 | Current high-volume, latency-sensitive tier for classification, extraction, and routing |