Anthropic has released Claude Haiku 5.5, a smaller model that the company says raises its computer-use test score from 15.7 to 72.4 percent while cutting token prices by as much as 90 percent. The model is priced at 0.10 US dollars per million input tokens and 0.50 per million output tokens below 100,000 tokens, with higher rates above that threshold.
At the same time Anthropic shipped Claude Sonnet 5.5, a mid-tier model that runs about 30 percent faster than its predecessor and is aimed at routine agentic and office work. The company also launched Claude for Startups, offering one free year of Claude Team with up to five premium seats and 1,000 US dollars in API credits for eligible young companies.
The releases landed within days of OpenAI's GPT-6 and Mistral's Large 4, underscoring a pricing arms race among frontier labs. Analysts note that when two labs independently land on identical pricing to the cent, it signals a floor rather than coincidence.
