T

Typhoon Ocr 7b

Developed by scb10x
A vision-language model specifically designed for Thai-English real-world document parsing, based on the Qwen2.5-VL-Instruction framework
Downloads 126
Release Time : 5/14/2025

Model Overview

Specialized in OCR recognition and structured parsing of Thai-English documents, supporting complex layout processing and multi-layer visual analysis

Model Features

Thai-English Bilingual Support
Specifically optimized for recognizing mixed Thai and English documents
Complex Document Parsing
Capable of processing structured documents like financial reports and government forms, as well as complex layout documents such as receipts and menus
Multi-layer Visual Analysis
Supports element recognition, contextual analysis, text extraction, artistic structure analysis, and comprehensive summary generation
Structured Output
Output supports Markdown, HTML tables, and <figure> tags, preserving the original document structure

Model Capabilities

Thai-English Bilingual OCR Recognition
Document Structured Parsing
Table Data Extraction
Chart Analysis
Multilingual Mixed Content Processing
Complex Layout Document Understanding

Use Cases

Financial Document Processing
Financial Report Parsing
Extracts structured data from complex financial reports
Outperforms GPT-4o and Gemini 2.5 Flash in performance
Government Document Processing
Government Form Parsing
Automatically identifies and extracts key information from government forms
High-precision structured output
Educational Material Processing
Academic Paper Parsing
Extracts text, charts, and reference information from academic papers
Supports Markdown and HTML format output
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase