Anthropic has released Claude Haiku 5.5, a compact model that starts at $0.10 per million input tokens. The pricing positions it as a low-cost option in the company's lineup, aimed at developers who need affordable inference.
Despite its small size, the model keeps a 1M-token context window, allowing it to process large documents or long conversations in a single pass. According to the release, it also achieves 72.4% on the OSWorld benchmark.
The combination of a low entry price, long context, and competitive benchmark performance suggests Anthropic is targeting high-volume use cases where cost and context length are critical. The model's exact availability and any additional usage tiers were not detailed in the announcement.