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

Anthropic has launched Claude Haiku 5.5, completing the company’s three-model Claude 5.5 lineup after the releases of Opus 5.5 and Sonnet 5.5. The new model is designed for high-volume and cost-sensitive workloads such as summarization, classification, information extraction, database queries and content compaction. Anthropic describes Haiku 5.5 as its fastest and most capable small model, making it particularly suitable for applications where speed and operating costs are important.

One of the biggest changes is its pricing. Anthropic says Haiku 5.5 costs around 75% less to run than its predecessor, Claude Haiku 4.5. For prompts under 100,000 tokens, the model is priced at $0.10 per million input tokens and $0.50 per million output tokens. Longer prompts carry higher rates. The lower cost is intended to make the model practical for large-scale deployments where businesses may process millions of requests.

Haiku 5.5 also brings capabilities aimed at real-time applications. Anthropic says the model works particularly well for live customer support, browser use, voice agents and in-app assistants, while also functioning as a subagent alongside larger Claude models on coding projects. The model supports a 1-million-token context window and up to 128,000 output tokens, with adaptive thinking that allows it to adjust how much reasoning it uses for a task.

Security is another part of the release. Anthropic says Haiku 5.5 is its first Haiku model with built-in safeguards for a narrow category of high-risk cybersecurity requests. The company says these protections are designed to address potentially dangerous cyber-related use cases without significantly affecting ordinary applications. The launch comes as Anthropic is also expanding its Cyber Verification Program, giving vetted security teams access to more powerful models for authorized vulnerability research and defensive work.

The release strengthens Anthropic’s position in an increasingly competitive AI market, where companies are competing not only on intelligence but also on inference costs and speed. Haiku 5.5 gives Anthropic a lower-cost option for everyday, high-volume workloads while Opus 5.5 and Sonnet 5.5 target more demanding applications. The rapid expansion of the Claude 5.5 family also comes as Anthropic prepares for a potential initial public offering, making its ability to demonstrate sustained model development and commercial growth increasingly important to investors.