MODELS
Leading AI models, processed in the EU
Your admin decides which models each team can use. New models are added as they become available in EU regions.
| Model | Provider | Best for | EU regions | Input / 1M tokens | Output / 1M tokens |
|---|---|---|---|---|---|
| Claude Sonnet | Anthropic | Writing, analysis, long documents |
| $3.47 | $17.33 |
| Claude Haiku | Anthropic | Fast everyday tasks |
| $1.16 | $5.78 |
| Llama 70B | Meta | Low-cost everyday tasks |
| $0.76 | $0.76 |
| Mistral Large | Mistral AI | European model, many languages |
| $0.63 | $1.89 |
| Nova Pro | Amazon | Text, files and images |
| $0.84 | $3.36 |
| Nova Lite | Amazon | Quick, very low-cost answers |
| $0.06 | $0.25 |
| Nova Micro | Amazon | Short text tasks, lowest cost |
| $0.04 | $0.15 |
| Cohere Command R+ | Cohere | Search and company knowledge |
| $3.15 | $15.75 |
| Nova Canvas | Amazon | Image generation |
| $0.04 per image | |
| Nova Reel | Amazon | Short video generation |
| $0.08 per second of video | |
| Amazon Transcribe | Amazon | Meetings and voice input, Polish and English |
| $0.025 per minute of audio | |
Prices in USD, excluding VAT, charged from your AI credit. 1M tokens is about 750,000 words. Model names, regions and prices: [CONFIRM BEFORE LAUNCH]
Three ways to choose a model
Auto
ongate.ai picks the best model for each prompt by task type and availability.
Low-cost
The cheapest model that can do the task well. Ideal for everyday questions.
Manual
Choose the model yourself, or send one prompt to two or three models and compare.
Your data is not used for training
Model providers process your prompts in EU regions and do not store or learn from them. If one model is unavailable, requests move to another automatically.