DREAM is one model that both understands images (like CLIP) and makes images from text (like top text-to-image models).
C2LLM is a new family of code embedding models that helps computers find the right code faster and more accurately.