DEEPSEEK JANUS PRO

Summary from the TLDR AI newsletter

DeepSeek's image model got a big upgrade today. It is a unified text and image model - sometimes referred to as any-to-any. This means that it can both understand and generate images and text interleaved in a conversation. This is similar to the omni models from OpenAI and the Gemini suite of models from Google.