# High-fidelity image synthesis

OmniGen2 is a powerful and efficient unified multimodal model composed of a 3B vision-language model and a 4B diffusion model, supporting visual understanding, text-to-image generation, instruction-guided image editing, and context generation.

AnyLoRA is a text-to-image generation model fine-tuned based on Stable Diffusion v1.5, suitable for fields such as art creation.