Claude 3 Haiku

Provider: Anthropic US

No longer listed on OpenRouter since 2026-09-26. The figures below are the last recorded values.
Input price
$0.25
/ 1M tokens
Output price
$1.25
/ 1M tokens
Cache price
$0.03
/ 1M tokens (read)
Context
200K tokens
Max output 4K tokens

Overview

Anthropic's Claude 3 Haiku is a cost-optimized model designed for high-volume, efficient AI workloads. It offers a large context window and multimodal capabilities, making it suitable for tasks requiring quick processing and affordability.

Strengths

  • Extremely low input pricing at $0.25/M tokens for cost-sensitive applications.
  • Features a substantial 200,000-token context window for extensive document processing.
  • Supports vision capabilities, enabling multimodal input understanding and analysis.
  • Includes function-calling, facilitating integration with external tools and APIs.
  • Offers competitive output pricing at $1.25/M tokens for efficient responses.

Best use cases

  • Processing high volumes of text and images for data extraction or summarization.
  • Automating customer support interactions with large knowledge bases.
  • Analyzing long documents or codebases for quick insights and synthesis.
  • Integrating AI capabilities into applications requiring tool use via function-calling.
  • Content moderation or classification where speed and cost are critical.

Pricing position

Claude 3 Haiku is Anthropic's most cost-effective option among its listed 'Haiku' models, with input priced at $0.25/M tokens and output at $1.25/M tokens. It matches peers like GPT-5 Mini and Gemini 3.1 Flash Lite on input cost but is more expensive on output than Qwen2.5 VL 72B Instruct ($0.75/M) and DeepSeek V3.2 ($0.378/M).

Caveats

While offering a vast context, its 4096-token maximum output may require chunking for very lengthy summaries or complex generations. For tasks demanding extremely low output costs, models like DeepSeek V3.2 offer significantly cheaper output pricing per million tokens. This model does not include a free tier for initial development or testing.

Notes drafted by an LLM from the figures on this page (2026-05-21) and rewritten when the price changes.

Key facts

Max output
4K tokens
Tracked since
2026-05-08
Last updated
2026-09-26

Features

Image inputFunction calling

Price history

  • 2026-09-26 No longer listed

Monthly cost estimate

Scale Input (M) Output (M) Monthly
0.1M (hobby) 0.1 0.05 $0.09
1M (small app) 1 0.5 $0.88
10M (mid scale) 10 5 $8.75
100M (large) 100 50 $87.50

※ Cache and batch discounts not factored in. Detailed calculator

Official resources

Other models from Anthropic

Similar-priced models

Source: OpenRouter API