Claude Haiku 5.5 arrives with massive price cuts proving the AI pricing arms race is far from over

Anthropic released Claude Haiku 5.5, its fastest and most affordable small model yet, with prices for short prompts up to 90% lower than Haiku 4.5—around 75% cheaper overall. The model targets high-volume tasks like summarization, classification, and live customer support. Benchmarks show big gains: 1,620 on GDPval-AA v2.1 versus 735, 72.4% on OSWorld-2.1 versus 15.7%, and 39.2% on Terminal-Bench 4.0 versus zero. Analyzing firm Artificial Analysis ranks it the top small model but notes high token consumption: around 162,000 output tokens per task at max effort, roughly three times GPT-6 Luna, putting it below the Pareto frontier on cost efficiency. Its hallucination rate is 40%, better than Luna's 77%. Haiku 5.5 is the first Haiku-class model with adjustable reasoning levels. Anthropic also cut Sonnet 5.5 cache read prices by 50% and introduced monthly API credits for Max and Team subscribers. Haiku 5.5 is available on AWS, Google Cloud, and Microsoft Azure.

(Source：The Decoder)