However, the model is not finalized and lacks safety features, allowing for the generation of explicit content. MattVidPro urged OpenAI to address these issues before public release. Despite the success of ChatGPT, OpenAI's image generator has not seen the same level of success, but the upcoming Dall-E 3 may change that. The exact public launch date remains unclear, but it could potentially launch before the end of the year.
Key takeaways:
- OpenAI is reportedly testing a new version of its imaging model, possibly Dall-E 3, which is invite-only and being tested through a preview inside ChatGPT 4.
- Tech YouTuber MattVidPro has praised the new model as the best AI image generator he has seen, outperforming competitors like Midjourney and Stable Diffusion.
- The new model can generate images in multiple aspect ratios and has shown impressive results in generating text-based images and reproducing copyrighted materials.
- However, the model currently lacks safety features, allowing for the generation of explicit and violent content, and needs fine-tuning before public release.