DeepSeek appears to be twisting the knife.
U.S.-based AI stocks took a tumble on Monday following the release of the China-based DeepSeek AI chatbot. The new product from the Chinese tech startup offers a more affordable large language model (LLM), presenting a competitive alternative to OpenAI’s options like ChatGPT.
But DeepSeek wasn’t finished. By Monday afternoon, the company unveiled its latest innovation: Janus-Pro-7B, a cutting-edge AI image generation model.
DeepSeek claims its Janus-Pro-7B outperforms existing models such as OpenAI’s DALL-E and Stable Diffusion. In a bold move that mirrors its approach with DeepSeek-R1, the company has made Janus-Pro-7B free and open source.
Mashable Light Speed
Like DALL-E, a user can input text describing a photo or artwork, and DeepSeek’s Janus-Pro will provide the user with an AI-generated image. DeepSeek says that Janus-Pro can both analyze and generate images.
“Janus-Pro is a novel autoregressive framework that unifies multimodal understanding and generation,” the company said in a technical report of the model. “It addresses the limitations of previous approaches by decoupling visual encoding into separate pathways while still utilizing a single, unified transformer architecture for processing.”
DeepSeek provided AI-generated image examples of the improvements between its prior Janus model, which can be viewed below.
Credit: DeepSeek
Though fairly new to the space, DeepSeek is already positioning itself as a formidable disruptor in the AI race, no doubt leaving industry leaders scrambling to adapt.