Anthropic Launches Claude Haiku 5.5 as Its Fastest and Cheapest AI Model

Anthropic has launched Claude Haiku 5.5, its latest small AI model, positioning it as the company’s fastest, cheapest and most capable Haiku model to date.

The new model is designed primarily for high-volume and cost-sensitive applications, including summarisation, database queries, classification, customer support and browser-based tasks. Anthropic said Haiku 5.5 can also support coding workflows alongside its more advanced Opus 5.5 and Sonnet 5.5 models.

One of the key features of the new model is its lower operating cost. Anthropic claims that Haiku 5.5 costs around 75 per cent less to run than Haiku 4.5. The company has priced it at $0.10 per million input tokens and $0.50 per million output tokens, compared with $1 and $5 respectively for Haiku 4.5.

Haiku 5.5 is the third model in Anthropic’s Claude 5.5 family launched within a month, following Opus 5.5 and Sonnet 5.5. Its launch comes as competition among AI companies intensifies and Anthropic prepares to expand its offerings across different price and performance segments.

According to Anthropic, Haiku 5.5 has also delivered strong results on several AI benchmarks. The company said the model outperformed OpenAI’s GPT-6 Luna on tests covering knowledge work, computer use, agentic coding and visual reasoning.

The model is also the first Haiku-class system to offer an adjustable effort setting, allowing users to choose between faster, lower-cost responses and greater computational effort for more complex tasks.

Anthropic said Haiku 5.5 includes improved safety and alignment measures compared with Haiku 4.5, including stronger cybersecurity safeguards and reduced instances of behaviour that could lead to misuse.

With its combination of speed, lower cost and improved capabilities, the model is aimed at businesses handling large volumes of AI-powered tasks.