Claude Sonnet 4.5 (latest) (EU)

Claude Sonnet 4.5 is presented by the aggregator as a balanced mid-tier entry in Anthropic's Claude Sonnet line, aimed at coding, analysis, and agent-style workflows where reasoning quality and cost control both matter. The entry sits in a closed-weights family and is exposed through a regional EU SKU routed via a third-party gateway, while still carrying Anthropic as the listed creator on the model family. Intended use, as the aggregator frames it, centers on practical production tasks such as software engineering assistance, structured reasoning, and tool-mediated agent loops, with multimodal input accepted as text, images, and PDF documents for grounded analysis.

From a working-fit perspective, the model pairs a very large that quick-info value window with a sizable per-response output ceiling, making it suitable for long-document reasoning, multi-file code review, and extended agent traces that need to stay within one prompt. Reasoning control, tool use, structured output, temperature tuning, and attachment handling are all enabled, which supports both deterministic JSON-style pipelines and more exploratory agent orchestration. Cache-aware request design is also attractive here, since prompt caching is exposed with read and write rates that can meaningfully reduce repeated-context cost for iterative coding or analytical sessions.