Anthropic

Claude Haiku 4.5

Claude Haiku 4.5 is a Claude 4 model from Anthropic, released in 2025-10. It costs $1 / $5 per 1M, has a 200k-token context window, and is best for classification, extraction, cheap-production. Last verified 2026-04-19.

Try Claude Haiku 4.5 Docs Take the picker quiz

Spec sheet

Pricing

Input
$1 / 1M
Output
$5 / 1M
Cached input
$0.1 / 1M
Batch discount
50%
Free tier
No

Context & speed

Context window
200k tokens
Max output
8k tokens
Throughput
~160 tok/s
Time to first token
~350 ms
Speed tier
fast

Capabilities

Tool use
Yes
Structured output
Yes
Prompt caching
Yes
Extended thinking
No
Vision input
Yes
Audio in / out
No
Fine-tuning
No

Deployment

Open weights
No
On-prem
No
HIPAA eligible
Yes
Zero retention
Yes
Regions
us, eu

Estimated monthly cost

Assumes typical token shape: 2k input, 600 output per call. Prompt caching is excluded from these figures.

10k calls/mo
$50.00
per month
100k calls/mo
$500.00
per month
1M calls/mo
$5.0k
per month

When to use Claude Haiku 4.5

Sweet spot

  • classification
  • extraction
  • cheap production
  • high throughput
  • chatbot

Known trade-offs

  • weaker on complex reasoning vs Sonnet

Best use cases

Best LLM for Data extraction

Cheap + structured output + vision for scanned docs

Best LLM for Chatbot

Fast + cheap + great instruction following

Best LLM for Classification

Best instruction following in the cheap tier

Best LLM for Summarization

Great summaries for the price, 200k context

Works with

Anthropic SDKClaude CodeCursorAWS BedrockGoogle Vertex AIVercel AI SDKLangChainOpenRouter

Compare Claude Haiku 4.5 to other models

vs GPT-5.4 Minivs o4-minivs Grok 4.20vs GLM-5.1vs Gemini 3 Flashvs Qwen 3.6 Plus

FAQ — Claude Haiku 4.5

How much does Claude Haiku 4.5 cost?

Claude Haiku 4.5 costs $1 / $5 per 1M tokens on the Anthropic API. Cached input reads cost $0.1 per 1M, cutting the input bill by roughly 90% on repeat system prompts. The batch API offers a 50% discount for async workloads.

What is the context window of Claude Haiku 4.5?

Claude Haiku 4.5 has a 200k-token context window with up to 8k tokens of output. That's enough for long reports, extended chat histories, or structured document analysis.

Does Claude Haiku 4.5 have a free tier?

No — No free API tier; new accounts get starter credits.

Is Claude Haiku 4.5 HIPAA / EU / on-prem friendly?

Claude Haiku 4.5 is HIPAA-eligible, available in EU regions, and is API-only. Zero data retention is available for enterprise customers.

What is Claude Haiku 4.5 best for?

Claude Haiku 4.5 is best for classification, extraction, cheap production, high throughput. Trade-offs to be aware of: weaker on complex reasoning vs Sonnet.

Which tools and SDKs work with Claude Haiku 4.5?

Claude Haiku 4.5 integrates with Anthropic SDK, Claude Code, Cursor, AWS Bedrock, Google Vertex AI, Vercel AI SDK, LangChain, OpenRouter. Most major AI frameworks support it either natively or through OpenAI-compatible endpoints.