Natively Multimodal Models

Models that handle text, image, audio and video within a single architecture from training onward: shared representation across modalities, early fusion, and cross-modal generation. Excludes multimodal tasks as a field, which belong to 4.4, traditional computer vision, which belongs to 4.3, and deepfakes, which belong to 11.5.

• 4.3 Computer Vision and Video
• 4.4 Multimodal AI
• 11.5 Deepfakes, Authenticity and Watermarking

• 5.1 Foundation Models and General-Purpose AI
• 5.2 The Transformer and Beyond
• 5.3 Mamba, State-Space Models, MoE and Emerging Architectures
• 5.4 Pre-training and Scaling Laws
• 5.5 Fine-tuning, RLHF, DPO and ORPO
• 5.6 Context Protocols and Tool Use
• 5.7 Multi-Agent Orchestration
• 5.8 Open, Open-Weight and Closed Models
• 5.10 Hallucination and Structural Reliability
• 5.11 Sovereign and Localized Models