Anthropic has added a third model to its rapidly expanding Claude 5.5 family, launching Claude Haiku 5.5 as the company's latest push to make advanced AI faster and more affordable for businesses. The new model, announced on October 7, comes just weeks after Claude Opus 5.5 and Claude Sonnet 5.5, giving Anthropic a three-tier Claude 5.5 lineup designed for different types of AI workloads. Unlike the flagship Opus model, Haiku 5.5 is aimed at situations where companies need to process a large number of requests quickly and keep operating costs under control. Anthropic says the model is particularly suited to tasks such as summarisation, classification, database queries, information extraction, customer support and browser-based applications..
Claude Haiku 5.5: What Is New?
Anthropic describes Haiku 5.5 as its fastest, cheapest and most capable small model so far. The model is designed for repetitive and high-volume workloads where using a larger frontier model for every request may not be economical. It can also work alongside larger Claude models as a subagent, particularly for coding and other multi-step tasks. For businesses, that distinction could be important. Instead of sending every task to a premium model, an AI application can use a smaller model for routine work while reserving more powerful systems for complicated reasoning or decision-making..
75% Lower Cost Than Claude Haiku 4.5
One of the biggest highlights of the launch is pricing. Anthropic says Claude Haiku 5.5 costs around 75% less to run than Haiku 4.5 on average. Its listed API pricing starts at $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens. Anthropic has also reduced the cost of using its larger models. The company said it has halved the cache-read price for Sonnet 5.5, making the model around 20% cheaper for most agentic workloads. That suggests the competition in AI is increasingly shifting beyond raw model intelligence. Speed, inference cost and the ability to run AI agents at scale are becoming equally important..
Built for AI Agents and Enterprise Applications
Haiku 5.5 is particularly relevant to the growing AI-agent market. AI agents often need to make multiple model calls while completing a task. If every individual call uses an expensive model, the overall cost can quickly increase. A faster and cheaper model can therefore act as a supporting layer for larger AI systems. Anthropic says Haiku 5.5 can be used for tasks such as quick lookups, summaries, data extraction and other repetitive operations while Opus 5.5 or Sonnet 5.5 handles more complex parts of a workflow. This could make the model attractive for enterprise software, coding assistants, customer-service systems and applications that need to process large volumes of information..
How Does Claude Haiku 5.5 Perform?
Despite its lower price, Anthropic is positioning Haiku 5.5 as a significant performance upgrade over its predecessor. On Anthropic's published benchmarks, Haiku 5.5 scored 72.4% on the OSWorld 2.1 computer-use evaluation, compared with 15.7% for Haiku 4.5. It also scored 39.2% on Terminal-Bench 4.0 for agentic coding, while Haiku 4.5 scored 0% on the same benchmark. The company also says Haiku 5.5 is the first Haiku model to include an adjustable effort setting, allowing developers to choose between lower cost and greater reasoning effort depending on the task. These are Anthropic's own benchmark results, so real-world performance can vary depending on the application, prompts and workload..
Claude 5.5 Family Is Now Complete
With Haiku 5.5, Anthropic now has three distinct models in its Claude 5.5 generation.
Claude Opus 5.5 is positioned as the flagship model for complex work and advanced reasoning. Anthropic launched it on September 22.
Claude Sonnet 5.5 arrived on September 28 and is designed as a faster and more economical option for everyday professional tasks, coding, documents and other workloads. Anthropic says it runs more than 30% faster than Sonnet 5 and can cost up to 30% less for most work.
Claude Haiku 5.5, the newest addition, focuses on speed, scale and affordability.
The three-model approach allows Anthropic to target different customers rather than forcing businesses to use one model for every application..
Why This Launch Matters in the AI Race
The latest launch comes as competition among leading AI companies is becoming increasingly intense. The focus is no longer simply on which company has the most powerful model. AI developers are also competing on cost per request, response speed, coding capabilities, agent performance and enterprise adoption. Anthropic's latest strategy reflects that shift. By expanding the Claude 5.5 range, the company is attempting to offer businesses a model for almost every layer of an AI workflow—from complex reasoning to routine high-volume operations. The timing is also significant. Anthropic has been accelerating its model releases, with Opus 5.5, Sonnet 5.5 and now Haiku 5.5 arriving within a matter of weeks. Reuters reported that the company launched Haiku 5.5 as its third Claude 5.5 model in roughly a month..
What It Means for Businesses and Developers
For businesses building AI-powered products, lower inference costs can have a direct impact on the economics of running those applications. Customer-support platforms, search systems, document-processing tools and AI agents can generate thousands or even millions of model requests. A cheaper model can therefore reduce operating expenses without requiring every task to be handled by a flagship system. For developers, Haiku 5.5's ability to work as a subagent alongside larger Claude models could also make multi-model AI systems more practical. In simple terms, the bigger model can think through the difficult part while Haiku handles the faster, repetitive work..
Anthropic's Bigger AI Strategy
The Haiku 5.5 launch shows that Anthropic is building Claude as more than a single chatbot. The company is increasingly positioning its models as infrastructure for enterprise software, coding tools and autonomous AI agents. Anthropic has also been expanding access to its advanced models for cybersecurity professionals through its Cyber Verification Program, showing how its latest models are being targeted at specialised professional use cases as well. As AI adoption moves from experimentation toward large-scale deployment, the ability to deliver powerful models at a sustainable cost could become just as important as benchmark performance. Claude Haiku 5.5 is not designed to replace Anthropic's flagship models. It is designed to make AI cheaper and faster to deploy at scale. With significantly lower operating costs, faster processing and improved performance across several benchmarks, Haiku 5.5 could become an important component of AI applications that handle large volumes of routine tasks. For Anthropic, meanwhile, the launch completes the initial Claude 5.5 lineup and strengthens its position in an increasingly competitive AI market where performance, speed and price are all becoming battlegrounds..
Source: Anthropic, Reuters. Performance figures and pricing are based on information published by Anthropic and may vary in real-world use..