Anthropic launches Claude Haiku 5.5 with lower costs and faster performance

Anthropic has launched Claude Haiku 5.5, the third model in its Claude 5.5 family, expanding its lineup with a model designed for high-volume, cost-sensitive AI workloads.

The company positions Haiku 5.5 as its fastest model to date and says it costs around 75% less to run on average than Haiku 4.5. The model is aimed at repetitive and speed-sensitive tasks including summarisation, classification, database queries, live customer support and browser use.

The launch comes as Anthropic continues to expand its model portfolio and prepares for a potential public offering, adding another product milestone to the company's rapid development cycle. Reuters reported that Haiku 5.5 is the third Claude 5.5 model released within a month.

A smaller model built for high-volume workloads

Haiku 5.5 is designed to handle tasks where speed, cost and scale are particularly important.

Anthropic says the model can support quick and repetitive workloads such as summaries, database queries and classification, while also serving as a subagent for coding tasks involving its larger Claude models.

The company also highlights applications such as live customer support and browser use, where faster inference can directly affect the user experience.

Lower pricing changes the economics of AI workloads

Cost is a central part of the Haiku 5.5 proposition.

Anthropic says Haiku 5.5 costs around 75% less to run on average than Haiku 4.5. For prompts up to 100,000 tokens, its listed input price is $0.10 per million tokens and output is $0.50 per million tokens.

The pricing is aimed at making workloads that may have previously been too expensive to run at scale more practical, particularly where organisations need to process large volumes of relatively focused tasks.

Performance moves closer to larger models

Although Haiku is positioned as a smaller model, Anthropic says Haiku 5.5 delivers substantial performance improvements over Haiku 4.5 across several benchmarks.

The company reports improvements in knowledge work, computer use, multidisciplinary reasoning and agentic coding, while acknowledging that larger models such as Sonnet 5.5 and Opus 5.5 remain better suited to more complex agentic coding tasks.

This positions Haiku 5.5 less as a replacement for Anthropic's larger models and more as a lower-cost component that can handle narrower tasks within broader AI workflows.

More models for agentic AI

Anthropic is also positioning Haiku 5.5 for use as a subagent alongside larger models.

In an agentic workflow, a larger model can manage a complex task while a smaller model handles individual steps such as retrieving information, summarising content or performing repetitive operations.

That model architecture could allow enterprises and developers to distribute workloads according to the level of reasoning required, rather than relying on a single high-cost model for every task.

Anthropic also cuts Sonnet costs

The Haiku launch comes alongside a pricing change for Claude Sonnet 5.5.

Anthropic says it is cutting the price of Sonnet 5.5 cache reads by 50%, reducing the cost of the model on most agentic workloads by around 20%. The company is also introducing monthly API credits for Max and Team subscribers to encourage experimentation with tools, applications and agents.

Together, the changes suggest Anthropic is focusing not only on increasing model capability but also on making its models more economical to deploy at scale.

A new model amid Anthropic's rapid expansion

Haiku 5.5 arrives during a period of rapid expansion for Anthropic's Claude portfolio.

The company has released three Claude 5.5 models within a short period, with Haiku following the launches of larger models in the same family. Reuters reported that the company is also preparing for a planned IPO, placing the continued expansion of its product portfolio in a broader period of corporate growth.

The launch also comes against a wider debate about the pace of frontier AI development. Anthropic CEO Dario Amodei has separately called for greater caution around the pace of frontier model releases, citing safety considerations.

The economics of scaling AI

Haiku 5.5 highlights an increasingly important part of the AI market: capability is only one measure of model progress.

For businesses deploying AI across thousands or millions of interactions, the cost and latency of each task can determine whether a use case is commercially viable.

By combining lower pricing with higher performance and faster inference, Anthropic is targeting the part of the market where AI systems need to operate repeatedly and at significant scale.

Haiku 5.5 is now available across Anthropic's platforms, as well as through Amazon Web Services, Google Cloud and Microsoft Azure.

About Anthropic

Anthropic is an AI company focused on developing reliable and interpretable AI systems. Its Claude family includes models designed for applications spanning knowledge work, coding, reasoning, agentic workflows and enterprise use.

Source and Credits

Anthropic's official Claude Haiku 5.5 announcement.