Model
Claude Haiku 4.5
Fast, cost-efficient Claude model for high-volume classification and extraction.
About
Overview
Claude Haiku 4.5 is built for speed and scale. Use it for classification, extraction, lightweight RAG, and routing — anywhere you need Claude-quality output at the lowest per-token price. Runs on Narrative Cloud via Amazon Bedrock.
Vendor Support
Specifications
- Family
- Claude
- Tier
- Fast
- Context Window
- 200K tokens
- Max Output
- 8K tokens
- Input Pricing
- $1 / MTok
- Output Pricing
- $5 / MTok
- Knowledge Cutoff
- 2026-01
- Modalities
- Text in, Image in, Text out
Capabilities
- Tool Use
- Supported
- Extended Thinking
- Not supported
- Prompt Caching
- Supported
- Batch API
- Supported
- Streaming
- Supported
- Vision
- Supported
Data Plane Dependencies
This model runs on the data plane providers listed below. Installing at least one makes the model available to use.
