Artificial intelligence firm Anthropic on Wednesday released Claude Haiku 4.5, a smaller version of its large language model designed to deliver faster responses and lower costs while maintaining high coding performance.
The company said Haiku 4.5 performs at roughly the same level as its earlier Claude Sonnet 4 model but operates at one-third the cost and more than twice the speed. The model is optimized for applications that require quick interaction, including chat assistants, customer support tools and code generation.
Anthropic said safety testing showed Haiku 4.5 had fewer alignment issues than previous releases and limited risk in producing harmful content. It is classified under the company’s AI Safety Level 2 standard, allowing broader use than higher-tier models.
The model is available through Anthropic’s API as well as on Amazon Bedrock and Google Cloud’s Vertex AI, with pricing set at $1 per million input tokens and $5 per million output tokens.