Anthropic launches Claude Haiku 5.5, its fastest and cheapest AI model yet

AI giant Anthropic has launched Claude Haiku 5.5, the third model in its Claude 5.5 family, as the company continues to broaden its artificial intelligence portfolio ahead of a planned initial public offering (IPO). This latest release by Anthropic follows the launch of Claude Opus 5.5 and Claude Sonet 5.5, completing a three-tier lineup aimed at different enterprise and developer use cases. Launching the new model Anthropic said that it is designed for high-volume, cost-sensitive workloads and is suited for tasks such as classification, summarisation, information extraction, live customer support, voice agents and in-app assistants. According to Anthropic, Haiku 5.5 is its fastest small model to date and is intended for applications where speed and affordability are critical.Anthropic is positioning Haiku 5.5 as a lower-cost alternativeAnthropic is positioning Haiku 5.5 as a lower-cost alternative within its model family. Reuters reported that the model costs about 75% less than its predecessor, Claude Haiku 4.5. For prompts under 100,000 tokens, pricing starts at $0.10 per million input tokens and $0.50 per million output tokens. For longer prompts, costs rise to $0.50 per million input tokens and $2.50 per million output tokens.The company said the model is built to handle repetitive and large-scale workloads efficiently, including database queries, summaries and classification requests. Anthropic also noted that Haiku 5.5 can work alongside its larger models, Opus 5.5 and Sonnet 5.5, as a sub-agent for coding and other complex workflows.New safeguards for cybersecurity-related requestsThe launch also introduces new safety features. Anthropic said Haiku 5.5 is the first Haiku-class model to include built-in safeguards for a narrow category of high-risk cybersecurity requests. The company added that these protections are designed to address specific security concerns while ensuring that ordinary day-to-day use cases remain unaffected.In addition, the model includes an adjustable effort setting that allows users to balance performance, intelligence and operating costs depending on their requirements. Haiku 5.5 supports a context window of up to one million tokens and can generate outputs of up to 128,000 tokens, according to Anthropic's documentation.Part of broader strategy before IPOThe latest launch comes as Anthropic accelerates product development ahead of a planned IPO. The company has been steadily expanding its AI offerings and enterprise customer base as competition intensifies among leading AI developers. Alongside the Haiku 5.5 release, Anthropic said it is reducing cache-read prices for Sonnet 5.5 and introducing monthly API credits for Claude Max and Team subscribers, moves aimed at making its platform more attractive to developers and businesses.You use AI every day. Now get your AI Quotient. Take the AIQ test.