According to a report by MarkTechPost, Anthropic has released Claude Haiku 5.5, a new small artificial intelligence model priced at $0.10 per million input tokens. The model features a one million token context window and achieved a score of 72.4 percent on the OSWorld benchmark.
Capabilities and Context Length
Small language models with extended context windows have become increasingly important for handling large volumes of enterprise data, document analysis, and complex coding tasks efficiently. By maintaining a one-million-token capacity at a low per-token cost, the release aims to make long-context processing more economically viable for developers. However, the exact pricing for output tokens was not specified in the available information, nor were the specific hardware requirements or inference speeds.
Unknowns and Future Performance
While the OSWorld benchmark score of 72.4 percent provides a quantitative metric for the model’s agentic and computer-use capabilities, real-world performance across diverse enterprise tasks remains to be seen. The announcement did not specify a precise rollout timeline, availability tiers for enterprise users, or the exact parameter count of the model. Observers will need to monitor independent benchmarks and developer adoption to gauge how Claude Haiku 5.5 compares to competing small-scale models in production environments.
Further details regarding rate limits and regional availability are expected to emerge as the model deploys more broadly.
Lead image: Brett Sayles / Pexels
