Anthropic launched Claude Haiku 5.5, its fastest and most capable small model, on October 8. It reduces costs by 75% versus Haiku 4.5, with input and output priced at $0.1 and $0.5 per million tokens respectively. Key features include adjustable reasoning intensity for cost-effective scaling. Concurrently, Anthropic halved Sonnet 5.5 cache read prices to $0.1 per million tokens, lowering operational costs for agent tasks by approximately 20%. This strategic pricing aims to optimize high-frequency, cost-sensitive scenarios like summaries and customer service while enhancing the efficiency of autonomous AI agents.