The Stable Video Diffusion family comprises AI models for video and 3D content generation, enabling image-to-video conversion, orbital video creation, and novel view synthesis through temporal diffusion architectures.

AuraFlow is an open-source text-to-image model family using Multimodal Diffusion Implicit Transformation architecture. Notable for flow-based generation and efficient scaling, it produces high-quality images up to 1536px with strong instruction-following capabilities.

MiniMax is a family of large language models featuring exceptional context length (up to 4M tokens), efficient parameter usage (456B total, 45.9B active), and both text-only and multimodal capabilities using hybrid attention mechanisms.

Whisper is OpenAI's family of speech recognition models using transformer encoder-decoder architecture, ranging from 39M to 1.55B parameters, capable of robust multilingual transcription, translation, and language identification across diverse acoustic conditions.