Anthropic Unveils Claude Haiku 5.5 as AI Race Intensifies

The launch, announced Wednesday, makes Haiku 5.5 the third model in Anthropic's Claude 5.5 family introduced over the past month, underscoring the company's push to broaden its AI portfolio as it prepares for a planned initial public offering. Claude Haiku 5.5 is designed for tasks that require fast and efficient processing rather than the most complex forms of AI reasoning. Anthropic said the model is suited to applications including classification, summarisation and information extraction.

The company also highlighted use cases such as live customer support, voice agents and in-app assistants, where businesses may need AI systems capable of handling large volumes of requests at relatively low cost. Anthropic cuts Haiku pricing One of the key features of the new model is its pricing. Anthropic said Claude Haiku 5.5 will cost $0.10 per million input tokens and $0.50 per million output tokens for prompts containing fewer than 100,000 tokens.

For longer prompts, the price rises to $0.50 per million input tokens and $2.50 per million output tokens. The pricing represents a 75 per cent reduction compared with the company's previous Haiku 4.5 model, positioning the new release as a more affordable option for developers and businesses seeking to integrate AI into products and customer-facing services. Lower inference costs are increasingly important as companies move beyond experimental AI deployments and begin incorporating models into everyday business operations, where thousands or millions of individual interactions can quickly increase computing expenses.

New safeguards target high-risk cybersecurity requests Anthropic also said Haiku 5.5 is the company's first Haiku model to include built-in safeguards designed to address a narrow category of high-risk cybersecurity requests. The company said the safeguards are intended to reduce the risk of misuse in sensitive cybersecurity scenarios, while noting that most everyday applications and tasks would not be affected. The move reflects the growing emphasis among AI developers on balancing model capabilities with safeguards against potentially harmful uses.

For businesses deploying AI in customer support, software applications and other routine functions, Anthropic said the new protections would not materially interfere with ordinary use cases. Third Claude 5.5 model in a month The launch adds another model to Anthropic's rapidly expanding Claude 5.5 family. With Haiku 5.5, the company is broadening the range of models available to customers, particularly those looking for an option that combines lower operating costs with the speed required for real-time applications.

The latest release also comes as competition among leading AI companies intensifies, with developers increasingly competing not only on model intelligence but also on speed, reliability, safety and the cost of running AI systems at scale. For Anthropic, the launch represents another step in positioning Claude as a broader commercial AI platform ahead of its planned IPO, with Haiku 5.5 targeting one of the fastest-growing segments of the market: affordable, high-volume AI deployment.