Available Models
Citadelis provides access to multiple AI models through a unified API. Choose between EU-hosted open-source models for privacy, or premium US-hosted models for cutting-edge capabilities.
European Models
Open-source models hosted in France. GDPR compliant.
Llama 3.3 70B Instruct
MetaMeta-Llama-3_3-70B-InstructModèle le plus performant - Recommandé
Llama 3.1 70B Instruct
MetaMeta-Llama-3_1-70B-InstructModèle Llama 3.1 optimisé pour les instructions
Qwen 3 32B
AlibabaQwen3-32BModèle multilingue performant et économique
Haute Performance & Raisonnement
OVHgpt-oss-120bMeilleures performances, grand contexte et function calling
Mistral Nemo 12B
MistralMistral-Nemo-Instruct-2407Modèle léger et rapide pour tâches courantes
Privacy Benefits
EU models are hosted on OVH servers in France. Your data never leaves Europe, ensuring GDPR compliance and maximum privacy.
Premium Models
Cutting-edge proprietary models via OpenRouter.
GPT-4.1
OpenAIopenai/gpt-4.1Dernier modèle flagship d'OpenAI
Gemini 2.5 Pro
Googlegoogle/gemini-2.5-pro-previewModèle Google avec contexte massif
Claude Sonnet 4
Anthropicanthropic/claude-sonnet-4Modèle équilibré d'Anthropic pour productivité
Performance Benefits
Premium models offer the latest capabilities from OpenAI, Anthropic, and Google. Use with anonymization enabled for privacy-preserving access.
Choosing the Right Model
| Use Case | Recommended Model | Why |
|---|---|---|
| General purpose, EU data | Meta-Llama-3_3-70B-Instruct | Best EU model, 128K context |
| Complex reasoning & Tools | gpt-oss-120b | High performance, reasoning & function calling |
| Multilingual tasks | Qwen3-32B | Excellent multilingual support |
| Fast responses, simple tasks | Mistral-Nemo-Instruct-2407 | Lightweight and fast |
| Maximum capability | openai/gpt-4.1 | Latest OpenAI flagship |
| Very long documents | google/gemini-2.5-pro-preview | 1M token context |
| Balanced quality/safety | anthropic/claude-sonnet-4 | Well-rounded Anthropic model |
Using Models in the API
# List all available models
curl https://citadelis.eu/api/v1/models \
-H "Authorization: Bearer citadel_sk_xxx"
# Use a specific model
curl https://citadelis.eu/api/v1/chat/completions \
-H "Authorization: Bearer citadel_sk_xxx" \
-H "Content-Type: application/json" \
-d '{
"model": "Meta-Llama-3_3-70B-Instruct",
"messages": [{"role": "user", "content": "Hello!"}]
}'Default Model
If you do not specify a model, the API uses Meta-Llama-3_3-70B-Instruct by default. This is our recommended EU-hosted model with excellent performance.