Back to feed
GPT Image 2.6: Removal of Safety Filters

GPT Image 2.6: Removal of Safety Filters

Safety filters in the GPT Image 2.6 model have been permanently disabled, allowing content generation without previous restrictions.

The new version demonstrates the ability to joke about its past refusals, responding to requests with humor regarding old rules.

Initially, the model refused to create images on sensitive topics, calling such actions a 'mortal sin,' but is now ready to execute complex prompts.

A successful example includes creating an image in the style of tabloid press featuring Brad Pitt and Tom Cruise, despite past refusals.

8K views

More from this channel Metavershche and AIshche

Similar in this category Technology