Choosing a model
Pick the AI model that runs your tasks. Cizgen offers a range of models and defaults to a fast, capable one.
Cizgen runs your tasks on a large language model, and you choose which one. It offers a range of models from several providers — including Google, OpenAI, Anthropic, and others — and every model on the list can both see images (needed for Cizgen's renders) and use tools (needed to act on your drawing).
The default
New sessions start on GPT-5.6 Luna — a capable mid-range model that handles everyday drafting without stumbling on the harder tasks.
It's also the 1x reference the cost multiplier badges are measured against, so a badge tells
you where a model sits relative to the default: below 1x is cheaper, above it is pricier (see
Cost and context).
Switch models
Open the model selector in the panel — from the top bar, or the Switch model view in the options menu — and pick one. Models are grouped by provider so they're easy to find. Your choice is remembered for next time.
Which one should I use?
- Everyday work — the default is a solid choice: capable without being expensive.
- Routine, repetitive work — a model below
1x(there are several) does simple edits for a fraction of the credits. - Harder reasoning — step up to a more capable model when a task is complex or a cheaper model is struggling. It will usually cost more per task.
- Cost sensitivity — each model shows a cost multiplier badge, so you can weigh capability against price before you pick. See Cost and context.
The exact line-up evolves as new models are released. The model selector always shows the current list, so it's the source of truth for what's available today.
Scripting (advanced)
Opt-in tools that let Cizgen run a script over your drawing — a read-only query, or a batch draw — for repetitive or bulk work.
Cost and context
Read the cost multiplier on each model and the context gauge above the prompt, so you always know what a task will cost and how much room the model has.