Language Models
AI-Public uses a central model catalog. The administrator sets, per provider, which quality levels are available; users then only see the models that are active according to that setting.
Providers and Quality Levels
For each provider, the environment can choose from:
| Setting | Effect |
|---|---|
| High quality | Users get the provider's stronger general model. |
| Cost-effective | Users get the provider's faster or cheaper model. |
| Combination | Users get both the cost-effective and high-quality model. Optionally, the top model can also be added. |
| Disabled | The provider and associated models are not available. |
The order of providers is determined by the environment. The default order is: OpenAI, Claude, Google, and European AI.
Current Catalog
| Provider | Cost-effective | High quality | Optional top model |
|---|---|---|---|
| OpenAI | GPT-5.6 Luna | GPT-5.6 Terra | GPT-5.6 Sol |
| Claude | Claude Haiku 4.5 | Claude Sonnet 5 | Claude Opus 5 |
| Gemini 3.5 Flash-Lite | Gemini 3.6 Flash | Gemini 3.1 Pro Preview | |
| European AI | Mistral Small 4 | Mistral Medium 3.5 | Mistral Large 3 |
Model Profiles
The model choice shows compact scores for speed, intelligence, and cost advantage. A higher score for cost advantage means the model is more cost-effective. A lower score thus means: more expensive to use.
| Model | Speed | Intelligence | Cost Advantage | Capabilities |
|---|---|---|---|---|
| GPT-5.6 Luna | 5 | 4 | 4 | Image input, reasoning and tools |
| GPT-5.6 Terra | 3 | 5 | 2 | Image input, reasoning and tools |
| GPT-5.6 Sol | 2 | 5 | 1 | Image input, reasoning and tools |
| Claude Haiku 4.5 | 5 | 3 | 4 | Image input, reasoning and tools |
| Claude Sonnet 5 | 3 | 5 | 2 | Image input, reasoning and tools |
| Claude Opus 5 | 2 | 5 | 1 | Image input, reasoning and tools |
| Gemini 3.5 Flash-Lite | 5 | 3 | 5 | Image input, reasoning and tools |
| Gemini 3.6 Flash | 4 | 4 | 3 | Image input, reasoning and tools |
| Gemini 3.1 Pro Preview | 3 | 5 | 2 | Image input, reasoning and tools |
| Mistral Small 4 | 4 | 3 | 5 | Image input, reasoning and tools |
| Mistral Medium 3.5 | 3 | 4 | 4 | Image input, reasoning and tools |
| Mistral Large 3 | 2 | 5 | 1 | Image input and tools |
Thinking Mode
When a provider is set to Combination, AI-Public can combine a fast model with a stronger model. The user can switch between them via the thinking mode:
| Mode | Model Type | Suitable For |
|---|---|---|
| Off | Cost-effective model | Quick questions, summaries, and daily tasks |
| On | High quality or top model | Complex analysis, long documents, and careful reasoning |
Use thinking mode especially when quality is more important than speed or cost.
