The models behind pictures, videos and vectors.
Pictures
Every picture, post and story is made with GPT Image 2.6 or Nano Banana 2.1. Settings lists both, ticked to begin with, and Cutlin picks the best ticked one for each request. Speed is set in Configuration: Auto, which is Fast unless the prompt asks for the most detail, Fast at 5 credits a picture, or Quality at 15, whatever the picture's size. Every plan includes every model the studio offers.
Videos
Kling 4.0 and Gemini Omni 1.2 make videos and change existing ones, and so do Seedance 2.5 and Kling O3 on Higgsfield. Cutlin picks the model from the video models ticked in Settings, and you choose what a video is for: Realistic, Cinematic, Cartoonish, Talking people, Big scenes or Quick clips. Auto leaves the look to your prompt.
Kling runs from 3 to 30 seconds. Gemini Omni runs from 3 to 15 and can change a video in conversation, turn by turn.
Higgsfield adds Seedance 2.5, Kling 4.0, Kling 3.0, Kling O3, Wan 3.0, MiniMax H3, Happy Horse 1.1, PixVerse V6, LTX-2.5, Grok Imagine 1.5 and Cinema Studio 4.0. Each has its own lengths and shapes, and each costs the same credits.
Vectors, and seeing what was used
Vector graphics and picture-to-SVG conversion use Recraft vector models. Captions and written answers come from GPT-6 Luna, and brands from GPT-6.1 Sol, with nothing to choose. Every result has a details view that lists each step that made it - reading the prompt, then making the picture or designing the brand - with its model, its tokens and its credits. Every AI result costs at least 1 credit; only the watermark is free.
