Use open models through an EU endpoint
Munito AI serves DeepSeek, GLM and Qwen models through an EU inference API.
Requests to https://api.eu.munito.ai/inference/v1 run in the EU.
Use the same OpenAI client library with a Munito API key and the chosen model ID.
The model developer and the inference provider have different roles. Munito AI operates the API that processes your request. Use the subprocessor list to assess the suppliers that can process your data.
Which model should you evaluate?
| Model | Model ID | Evaluation focus | Details and current EU prices |
|---|---|---|---|
| DeepSeek-V4-Flash | deepseek-ai/deepseek-v4-flash | Text tasks, code and reasoning | DeepSeek API |
| GLM-5.3-Flash | zai-org/glm-5.3-flash | Text, images, tools and structured outputs | GLM API |
| Qwen3.8-27B | qwen/qwen3.8-27b | Text, images, document pages and tools | Qwen API |
Choose a model for your task, then compare it on your own evaluation set. Each model page lists context limits, capabilities and available regions. Use the full catalog for other versions and model families.
Compare costs for your workload
Select the EU region in the catalog. Compare input, cached input and output prices for the same service tier. Include reasoning tokens in the output budget when the model uses them. A lower input price alone does not establish a lower cost per completed task.
Use a fixed evaluation set to measure answer quality, latency and token use. Test the concurrency that your application requires. The rate limit guide explains capacity responses and account limits.
Are prompts sent to the model developer?
Munito AI serves the open model through its own API. The name of a model does not mean that a request goes to its developer's hosted API. The published subprocessor list describes the service's suppliers. Optional overflow can use another provider or model when you enable it.
What stays in the EU?
The EU endpoint constrains inference to the EU. The EU data region keeps stored customer content in the EU. Run on managed European infrastructure, inside your own cloud account, or on-premises.
Munito AI keeps no prompts or outputs by default and does not train on your data. Explicit storage and background requests have separate retention rules. Read the EU API guide for deployment choices and integration steps.
Start an evaluation
- Create an account and choose the EU data region.
- Create an API key in the console.
- Set the EU base URL and the model ID.
- Run synthetic examples from your application.
- Check quality, latency, token use and the security requirements of your customer.
Follow the quickstart for code examples. Use the security review guide for procurement questions. For capacity requirements, talk to an engineer.