Skip to content

Models

Open-weight and frontier in one place. No limits on model choice.

Use frontier and open-weight models side by side, each for the tasks it handles best. We’ll work out with you which models fit your company.

  • Frontier models
  • Open-weight models
  • Your own models
  • EU region
  • Germany

OpenAI

Frontier and open-weight

  • GPT-6 AstraHosted in the EU
  • GPT-6.1 SolHosted in the EU
  • GPT-6.1 TerraHosted in the EU
  • GPT-6 LunaHosted in the EU
  • gpt-oss-120bHosted in Germany

Anthropic

Frontier

  • Claude Opus 5.5Hosted in the EU
  • Claude Sonnet 5.5Hosted in the EU
  • Claude Haiku 5.5Hosted in the EU

xAI

Frontier

  • Grok 4.7Hosted in the EU

DeepSeek

Open-weight

  • DeepSeek 4.1Hosted in Germany

Qwen

Open-weight

  • Qwen 3.8 27BHosted in Germany
  • Qwen 3.5 9BHosted in Germany

Mistral AI

Open-weight

  • Mistral Small 24BHosted in Germany

Moonshot AI

Open-weight

  • Kimi K3Hosted in the EU

Zhipu AI

Open-weight

  • GLM 5.3Hosted in the EU

NVIDIA

Open-weight

  • NemotronHosted in the EU

Your own model

Any provider

We connect your own or fine-tuned models from your cloud or data center to Intra AI. They run right alongside all the other models.

Talk to us

More models are added regularly.

How to choose a model

The flag next to each model shows whether it is hosted in Germany or in the EU region. If you want your requests processed in Germany as well, pick a model with the German flag.

In a chat, you pick the model in the header, and you can switch it in the middle of a conversation. The model picker shows for each model whether it thinks before answering, whether it reads images and, where stated, how large its context is. Every answer is labeled with the model that wrote it.

Your administrators set up which models are available on your instance. An agent can also fix its own model. We’ll work out with you in the intro call which models fit your tasks.