Claude Haiku 5.5

Your AI assistant tasks now cost up to 75% less if you switch to Claude Haiku 5.5, and specific agentic tasks are 20% cheaper on Sonnet 5.5.

Anthropic has released Claude Haiku 5.5, its smallest, cheapest, and fastest AI model to date. The company states Haiku 5.5 is designed for high-volume, cost-sensitive tasks such as summarization, database queries, and classification. It can also function as a subagent for coding tasks alongside larger models like Opus 5.5 and Sonnet 5.5. Anthropic reports that Haiku 5.5 costs approximately 75% less to run than its predecessor, Haiku 4.5, and offers improved performance across various benchmarks, including knowledge work, computer use, and multidisciplinary reasoning. The model also features an adjustable effort setting, allowing users to optimize for either cost or intelligence. Additionally, Anthropic has reduced the price of cache reads for Claude Sonnet 5.5 by 50%, making it about 20% cheaper for agentic work, and is introducing a new monthly API credit for Max and Team subscribers.

The introduction of Claude Haiku 5.5 marks a significant push by Anthropic to make advanced AI capabilities more accessible and economical for high-frequency, cost-sensitive applications. By drastically reducing the price and increasing the speed of its small model, Anthropic is targeting a market segment that previously found AI integration cost-prohibitive. This move could accelerate the adoption of AI agents for routine tasks across various industries, from customer support to data processing, potentially lowering operational costs for businesses and enabling new service offerings that rely on rapid AI responses.

Anthropic's strategic pricing adjustments, including the 50% cut on Sonnet 5.5 cache reads, signal a broader trend towards optimizing AI service delivery for efficiency and scale. The focus on lower-cost, faster models like Haiku 5.5 suggests a maturing AI market where specialization and cost-effectiveness are becoming key differentiators. This could lead to increased competition among AI providers to offer tiered models that cater to specific use cases, pushing the industry towards more granular pricing and performance trade-offs for developers and end-users.