Models & architecture
Multimodal model
A model able to process or generate more than one type of content.
Definition
A multimodal model works across combinations of text, images, audio, video or structured inputs within one model or coordinated system.
Why it matters
Bank documents mix prose, tables, charts, scans and signatures, so multimodal quality and data-handling controls require separate evaluation.
Related concepts
- Foundation model
A large pre-trained model adaptable to many downstream tasks.
- Document AI
AI used to classify, extract and reason over complex documents.
- System under test
The complete configuration being evaluated—not only the model.