Anthropic unveils Claude Haiku 5.5, its cheapest AI model with adjustable effort setting

Anthropic has announced the release of Claude Haiku 5.5, positioning it as the cheapest, fastest, and most capable small model in the company's lineup. The new model represents a significant price reduction from its predecessor, with Anthropic claiming it costs approximately 75% less to run on average than Haiku 4.5. The release comes as competition in the AI sector heats up, with multiple companies racing to deliver more affordable options for developers managing high-volume, cost-sensitive applications.

The model is specifically designed for tasks requiring speed and budget efficiency, including document summaries, database queries, classification requests, and live customer support operations. Notably, Haiku 5.5 is Anthropic's first Haiku-class model to feature an adjustable effort setting, giving users the ability to prioritize either cost savings or intelligence optimization depending on their use case. The company also highlighted its utility as a subagent when paired with the larger Claude Opus 5.5 and Claude Sonnet 5.5 models, particularly for coding workloads that benefit from tiered processing strategies.

Initial community reaction has been largely positive, with developers highlighting practical use cases such as fan-out checks, triage operations, and codebase scouting. However, some voices in the developer community have raised questions about quiet failure modes on complex editing tasks and whether reported benchmark gains reflect genuine capability improvements or adjustments to evaluation methodologies. As the AI pricing war continues to intensify across the industry, Haiku 5.5's arrival signals further pressure on competitors to match both performance and affordability standards.