Claude 3 Haiku
Claude 3 Haiku was the fastest and lowest-cost model in Anthropic's Claude 3 family, released in March 2024 for workloads that needed short response times and high throughput. It offered a 200,000-token context window and was available through the Claude API, Amazon Bedrock, and Google Cloud Vertex AI. Anthropic did not publish its architecture, parameter count, or inference optimizations; the documented training approach is constitutional AI, applied across the Claude 3 family. Anthropic announced the deprecation on February 19, 2026 and retired the model from the Claude API on April 20, 2026, after which requests to claude-3-haiku-20240307 return an error.
Anthropic did not disclose the compression or inference techniques behind the model's speed. Publicly, it was positioned for real-time applications, high-volume processing, and cost-sensitive deployments, under Anthropic's stated goal of helpful, harmless, and honest behavior. The model was used for conversational AI, content moderation, document processing, customer service automation, and light analytical work, where speed and cost per token mattered more than peak reasoning quality. Teams applied it to real-time customer support, automated content analysis, high-throughput document processing, and other cost-sensitive deployments, through direct API calls or batch jobs. Anthropic's recommended replacement for workloads still referencing it is Claude Haiku 4.5.
Related terms
Related services: Agentic AI consulting.
Ready to put agentic AI to work?
Book a free 45-minute consultation. We'll map one real process worth automating with production-grade AI.