Models and Limits
Which AI models you can choose, how automatic mode picks one, and how to spend your usage budget well
DeutschlandGPT brings the leading AI models together in one place. You choose per conversation which model answers. Every model runs on infrastructure in Europe, either at the European vendor or in a European cloud region.
Choose a Model
Set the active model from the model picker at the top of the input area. The switch takes effect from the next message; the conversation continues seamlessly. If you would rather not decide, leave Automatic selected.
Which Models Exist
Models from several vendors are available. Exactly which ones are enabled depends on your organization. The main families:
Very strong at demanding writing, analysis, and coding. Sonnet is the versatile default for everyday and demanding tasks alike.
Versatile all-rounders, plus reasoning models for multi-step, logical thinking.
Very fast and uses your limit slowly. The Flash variants are ideal for simple, frequent tasks.
Open and European models for standard tasks that go easy on your limit.
Models differ in their capabilities: not every one understands images, can use tools, or handles very long inputs. The model picker shows these traits right on each entry. Which tools a model can use also depends on the model.
Automatic Mode
Automatic makes the choice for you. It estimates the complexity of your request and routes it to a fitting model: a simple question to a fast model that uses your limit slowly, a demanding task to a stronger one. For most users this is the best default, because it delivers a good balance of quality and consumption without any thought.
Understanding Limits
Usage is measured in tokens, the text units that make up input and output. Each model uses your limit at a different rate: powerful models draw on it faster, fast models considerably slower.
Your organization sets a monthly limit. What happens once it is reached depends on the configuration:
- If the limit is reached, DeutschlandGPT can automatically switch you to leaner budget models so you can keep working.
- Alternatively, your organization can enable pay-as-you-go billing, so usage beyond the limit runs separately.
- On top of that, fair-use limits apply to absorb short bursts of load.
You can find your current usage and limit under Settings → Usage.
Use Your Limit Wisely
You get the most out of your limit by matching model strength to the task:
| Task | Recommendation |
|---|---|
| When in doubt, or for mixed work | Let Automatic choose |
| Demanding writing, analysis, coding, reasoning | A strong model like Claude Sonnet |
| Short questions, summaries, rewrites, routine work | A fast model like Gemini Flash that goes easy on your limit |
The simple rule of thumb: reach for a top-tier model only when a task truly needs it. For most day-to-day work a fast model is plenty and uses your limit far more slowly.