Sources
Loading...
Additional media
Loading...
DeepNewz, mobile.
People-sourced. AI-powered. Unbiased News.
A mysterious new text-to-image model, referred to as 'Blueberry', has appeared on the @ArtificialAnlys leaderboard, sparking speculation within the AI community. Some experts suggest that 'Blueberry' might be an advanced version of the Sora model, which has previously demonstrated impressive capabilities as seen in a technical report back in February. Others believe it could be related to GPT-4o's image generation abilities, originally named O1, or even a new version of DALL-E. Additionally, there is speculation about Midjourney 7, although it is unlikely as it lacks an API. The release of this model is anticipated to occur in October, potentially alongside other updates such as GPT-4.5 or Sora2.
New text-to-image "blueberry" models in the arena - thoughts? Maybe new SD model, GPT-4o image gen., or a new version of DALL-E 3? https://t.co/6Dqa0JyG93
Blueberry sounds like strawberry. It could indicate that we will get a revised dalle or Sora in October, because as The Information says, they are already working (training) on Sora2. I consider the release of a weaker Sora or a revised Dalle alongside GPT-4.5 to be extremely… https://t.co/kX50ylPZHR
What is 🫐 (blueberry)? Place your bets. Which image (or video) model could it be?: 1. Sora (old or new upcoming model) 2. Grok 3. GPT-4o 4. Midjourney 7 5. Mystic 2 6. Something else It won't be #4, MJ does not have an API, but it is coming in October or November, along… https://t.co/EMVKBt7bo6
