Prompt Caching with Claude: Faster, Cheaper AI Interactions
What is Prompt Caching?
Enables developers to cache frequently used context between API calls, providing Claude with more background knowledge while reducing costs and latency.
💰
Up to 90% cost reduction
⚡
Up to 85% latency reduction
When to Use Prompt Caching
- 🗨️ Conversational agents
- 💻 Coding assistants
- 📄 Large document processing
- 📝 Detailed instruction sets
- 🔍 Agentic search and tool use
- 📚 Talk to books and long-form content
Performance Improvements
| Use Case | Latency w/o Caching | Latency w/ Caching | Cost Reduction |
|---|---|---|---|
| Chat with a book | 11.5s | 2.4s (-79%) | -90% |
| Many-shot prompting | 1.6s | 1.1s (-31%) | -86% |
| Multi-turn conversation | ~10s | ~2.5s (-75%) | -53% |
Pricing for Cached Prompts
Claude 3.5 Sonnet
Our most intelligent model
- Input: $3 / MTok
- Cache write: $3.75 / MTok
- Cache read: $0.30 / MTok
- Output: $15 / MTok
Claude 3 Opus
Powerful model for complex tasks
- Input: $15 / MTok
- Cache write: $18.75 / MTok
- Cache read: $1.50 / MTok
- Output: $75 / MTok
Claude 3 Haiku
Fastest, most cost-effective model
- Input: $0.25 / MTok
- Cache write: $0.30 / MTok
- Cache read: $0.03 / MTok
- Output: $1.25 / MTok
Customer Spotlight: Notion
Notion is adding prompt caching to Claude-powered features for its AI assistant, Notion AI. This optimization allows for faster and cheaper operations while maintaining high-quality results.
"We're excited to use prompt caching to make Notion AI faster and cheaper, all while maintaining state-of-the-art quality."
— Simon Last, Co-founder at Notion
Get Started with Prompt Caching
Explore our documentation and pricing page to start using the prompt caching public beta on the Anthropic API.
@Anthropic
Tags：Prompt CachingClaude AIAPI cost reductionlatency reductionconversational agentscoding assistantsAI performance improvementNotion AI