clef: 27B Multimodal Model for Extracting Structured Data from Image Inputs

This is a multimodal model fine-tuned by Cloudflare based on Qwen3.8-27B. It accepts both image and text inputs to perform classification or extraction tasks, specializing in returning results as structured data in a predefined format.

Beyond simple text generation, it supports an 'image-text-to-typed-output' pipeline. It is suitable for automation workflows that require analyzing screen captures or document images to accurately extract specific field values, similar to scraping.

Distributed under the Apache 2.0 license, it allows for free commercial use. With a scale of 27B parameters, it maintains high inference performance while being useful for developers building image-based data parsing or document processing pipelines.