Perplexity has recently introduced two new image generation models for users with a Pro tier subscription. In the editor of Perplexity pages, users can generate images based on prompts that summarize the content of the article. Now, users have the option to select the Flux version 1 model as well. Flux version 1 is an open-source image generation model that performs on par with advanced models like Midjourney. This same model was recently introduced in Grok 2.
Perplexity also uses a development version of this model called "Schnell", which has shown promising results. In addition, Playground AI has upgraded its image generation model with the release of version 3 beta, now accessible on Perplexity. This version is also available on the Playground website, producing slightly different results.
Users now have the option to choose between:
- Playground v3
- Flux.1
- DALL·E 3 from OpenAI
- Stable Diffusion XL
Flux is currently considered the most advanced among these options. Additionally, a small UI change allows users to adjust prompts for image generation more precisely by specifying specific styles, which should help produce better results.