Wynfall models.
Your messages reach Sigel to answer you. That is the whole job.
One model,
Sigel.
Sigel is the Wynfall model: deeper reasoning for code, writing and long documents, with up to 262K tokens of context. It is on every plan, and Free uses Sigel with a small daily allowance. Its mark is three feathers.
Sigel is in early access On every plan
The models
- Good for
- Deeper reasoning for code, writing and long documents: work that crosses files, systems and decisions.
- Plans
- Every plan. Free uses Sigel with a small daily allowance, about 6 messages a day that refill at midnight Eastern. Starter, Builder and Studio add Wynfall Code, Standard, Extended or Max thinking and much more usage. When a paid plan runs out of usage, Sigel waits until it refills.
- Effort levels
- Low, Medium, High or Max.
Every model, the same rule.
Every Wynfall model keeps one rule. Nothing you give it trains a model, on any plan, Free included, unless you send it to us as feedback. The rule is written into our Terms.
Your own providers.
Connect an API key in Desktop or Code and that provider answers under its own terms and billing.
Local models.
Run a model on your own computer with Desktop or Code, and the model runs on your hardware.
Also in research.
Wynfall Image v1 is our first image model. Describe a picture in words and get product shots, characters and scenes back. It is in research and is not available yet.
