Anthropic has launched Claude Haiku 5.5, positioning it as the fastest and most cost-effective model in its current lineup. The company reports that the average operational cost for this model is approximately 75% lower than that of the previous Haiku 4.5 version.
The new model features a 1 million token context window and a maximum response size of 128,000 tokens. For requests under 100,000 tokens, the pricing is set at $0.10 per million input tokens and $0.50 per million output tokens. This represents a 90% reduction in token pricing compared to the predecessor.
Haiku 5.5 introduces a customizable effort level, allowing developers to balance reasoning quality against speed. In performance benchmarks, the model achieved 72.4% in OSWorld 2.1 and 39.2% in Terminal-Bench 4.0, showing significant improvements over the 15.7% and 0% scores of Haiku 4.5 respectively.
Anthropic also updated its pricing for the Sonnet 5.5 model, cutting the cost of cache reads by 50% from $0.20 to $0.10 per million tokens. This adjustment is expected to reduce overall operating costs for Sonnet 5.5 by roughly 20% in most agentic workflows.
The model is now accessible via the Claude Platform, Amazon Web Services, Google Cloud, and Microsoft Azure. Additionally, Anthropic is introducing monthly API credits for Max and Team plan subscribers to support the development of custom agents and applications.
