Docs / Administration / LLM hosting and regions
LLM hosting and regions
Compare models served through Berget AI’s EEA data centres, Mistral’s European endpoint and other providers used by OAIZ.
OAIZ offers three Berget AI models served through data centres in the European Economic Area (EEA). Mistral Large 4 uses Mistral’s European regional endpoint, which covers EU and EFTA countries.
Inference is the processing of your prompt to produce a model response. The regions below describe OAIZ’s default provider routes for that model call. Other workflow actions, connected services and storage have their own processing locations.
Berget AI in the EEA
Select the Berget AI version of a model to use Berget’s infrastructure. Its DPA states that its data centres are within the EEA, without a Sweden-only commitment.
| Model |
|---|
| GLM 5.3 Flash |
| Kimi K3.1 Pro |
| Gemma 4 31B |
Mistral in EU and EFTA countries
Mistral hosts Mistral Large 4 in EU and EFTA data centres through its European regional endpoint.
| Model |
|---|
| Mistral Large 4 |
Gemini through Google Cloud
Google Cloud hosts Gemini models through Vertex AI’s global service.
| Model |
|---|
| Gemini 4 Carbon |
| Gemini 3.8 Flash |
| Gemini 3.5 Flash Lite |
| Gemini 3.5 Flash |
| Gemini 3.1 Flash Lite |
| Gemini 3.1 Pro |
| Gemini 3 Flash |
OpenAI
OpenAI serves the following GPT models directly through its API.
Groq
Groq hosts OpenAI’s open-weight GPT OSS models on its own inference infrastructure.
| Model |
|---|
| GPT-OSS 120B |
| GPT-OSS 20B |
Cerebras
Cerebras serves GPT OSS, Gemma and Qwen models through its inference cloud.
| Model |
|---|
| Qwen 3.8 27B |
| Gemma 4 31B |
| GPT-OSS 120B |
Grok through xAI
xAI serves Grok models through its global API.
| Model |
|---|
| Grok 4.7 |
| Grok 4.6 |
| Grok 4.5 |
| Grok 4.3 |
| Grok 4.20 Reasoning |
| Grok 4.20 Non-Reasoning |
Anthropic
Anthropic hosts Claude on the Anthropic API. Default routing covers select countries in the United States, Europe, Asia, and Australia. Customer data is stored in the United States.
Other models through OpenRouter
Hosted By lists the providers OAIZ uses for each model through OpenRouter.
| Model | Hosted By |
|---|---|
| Muse Spark 1.3 | Meta |
| GLM 5.3 Flash | Wafer, Sail Research, Parasail, NextBit |
| GLM 5.3 | Novita, Together AI, Fireworks, Parasail |
| DeepSeek V4 Pro 0813 | Together AI, Cloudflare, Fireworks, DeepInfra, Parasail, GMICloud |
| DeepSeek V4 Flash 0731 | DeepInfra, Sail Research, Together AI, Parasail |
| Kimi K3 | Modal, Baseten, Fireworks, Together AI |
| GLM 5.2 | Novita |
| Kimi K2.7 Code | AtlasCloud, Parasail, Novita, Together AI |
| MiniMax M3 | AtlasCloud, Novita, DeepInfra, Parasail, Together AI |
| Kimi K2.6 | Novita, DeepInfra, Chutes, Parasail |
| GLM-5 | AtlasCloud, GMICloud, DeepInfra |
| Kimi K2.5 | AtlasCloud, Novita, Chutes, DeepInfra |
| DeepSeek V3.2 | AtlasCloud, Novita, DeepInfra, Parasail |
DeepSeek V4 Pro 0813 requests with structured output use Together AI, Cloudflare, Fireworks or Parasail.