Anthropic announced Claude Haiku 4.5, a small model the company says cuts cost and latency while delivering competitive coding performance.

Anthropic said Haiku 4.5 offers coding capabilities comparable to Claude Sonnet 4 at roughly one-third the price and more than double the speed, aimed at real-time and low-latency use.

The company highlighted that Haiku 4.5 not only rivals but in some tasks surpasses previous models, particularly in scenarios involving workflow orchestration and computer use.

In a customer quote, Andrew Filev, CEO of Zencoder, said, “Claude Haiku 4.5 is remarkably capable—just six months ago, this level of performance would have been state-of-the-art on our internal benchmarks.”

For safety and reliability, Anthropic reported comprehensive evaluations, noting that Haiku 4.5 demonstrated a lower incidence of misaligned behaviors than earlier versions, resulting in its classification under the less restrictive AI Safety Level 2.

Claude Haiku 4.5 is available on Anthropic’s API, Amazon Bedrock and Google Cloud’s Vertex AI, where Anthropic says it serves as a “drop-in replacement” for Haiku 3.5 and Sonnet 4.