Model overview
What this model is for
clef:27b is a 27B multimodal model from Cloudflare designed to turn a state and a schema of typed questions into decisions. Its documented modality is image input, and its primary practical use case is typed decision workflows that combine visual information with structured questions and outputs. This makes it potentially relevant for applications that need schema-driven decisions rather than general conversational generation. The source does not disclose a context-window size, inference speed, evaluation results, pricing, or a detailed list of supported capabilities beyond vision and decision-oriented use. Licensing terms are also not disclosed on the provided model page, so review the official source before deployment or redistribution. The model page reports 415 downloads and says it was updated 21 hours ago, making the listing relatively fresh according to the source. However, the page does not provide enough information to assess accuracy, latency, production readiness, safety behavior, or performance under specific workloads. Validate image handling, structured-output reliability, and decision quality against representative data before adopting it.