Next OpenAI image creation model multimarket
7
21
420
2025
74%
It will come out in 2024
66%
It will allow more than 8 aspect ratios to be generated directly
54%
It will allow resolution greater than 1792x1024 px in terms of megapixels
50%
It will be out not later than Q3 2024
50%
I will perceive it as more tolerant or better at generating pinups (I won't trade in this one)
50%
It will allow you to ask for music or sound effects to be generated along with it
50%
It will be available via API within first two weeks of paid user web demo availability
50%
Images will have effective watermarks such that even if you crop or scale, some sites can still tell it's an openAI image
50%
Cost per image of de3 equivalent images goes down (same megapixel, equivalent quality minimum)
50%
The best default image output from a prompt per image cost will be equal to or cheaper than de3's best cost by API (ie picking the largest size and best quality for both) comparing today's de3 and the future API cost then
39%
Censorship rules are loosened so that, say, USA's 1980s PG movie level nudity is allowed
34%
Long quotes of 30 words, if asked to be included as text in the image, come out letter-correctly at least 80% of the time. (i.e. the english is correct)

This is only about the image generation part directly. i.e. you input a prompt <or something else> and ask to generate an image out.

This is about the next paid or free major Image generation system release from OpenaAI, something like Dalle-4, or else the next one released, even if it has a new name. It has to be in broad release, enough for me, a paying user (up to ~200USD ppp equivalent) to use, per month.

Closes 2 weeks after availability is reached, based on the version at that point. (i.e. items can YES early, but can't NO until 2 weeks after the point of availability, so that initial bugs/issues/hardware limitations can be resolved).

Get Ṁ200 play money

More related questions