Model catalogue · Meta
Llama
Vendor version: llama-4-maverick (as of 29 August 2026)
A proven open text model from Meta for summarising, drafting and correspondence. Because the weights are openly available, there is no model vendor anything is transmitted to; on complex legal argument the text quality sits below the Claude tiers. Under professional-secrecy law this model is broadly positioned: in dedicated operation on reserved hardware in Germany, at a German EU operator, via the Open Telekom Cloud and via Azure on reserved capacity the chain is closed in each case - the element required by section 203 of the German Criminal Code, the confidentiality undertaking of the participating operator, is signed and on file. At IONOS, by contrast, the product-specific confirmation of that clause is still outstanding.
Everything section 203 requires is settled, including the provider's confidentiality undertaking. You do not have to clarify anything further.
The operator bills separately - input EUR 0.65 · output EUR 0.65. The figure above is the blend of the two, weighted three quarters input to one quarter output, as it arises in a firm. Passed through at purchase price, without a mark-up.
What it is good for
- Proven open text model for summarising and drafting
- Open weights - operation without a US cloud in the chain is possible
- Available free of CLOUD Act exposure through the Gosign chain - in the EU as well as in German data centres
Where it stops
- For multi-step analysis tasks, reasoning models are the better choice
- Knowledge ends with the training cut-off - current case law and facts belong in the context
- Citations and references always have to be verified professionally
- On the Telekom route the commitment covers only those models operated there on the T-Cloud in Germany - this one is among them; it does not apply to the GPT, Claude and Gemini lines passed through in the same catalogue (research date 2026-08-17)
Routes examined
The same model, procured differently, is a different legal position. What is examined per route is the contract chain, not the model.
| Route | Section 203 status | Version on this route |
|---|---|---|
| Dedicated instance in Germany, operated by Gosign | covered | - |
| EU inference of open models through the Gosign chain | covered | - |
| IONOS AI Model Hub (German data centres) | conditional | - |
| German data centres through the Gosign chain | covered | - |
| Microsoft Azure | covered | - |
The conditions behind a conditional status - which piece is open on which route and who closes it - are set out in the full matrix in the German catalogue, and we go through them in the conversation. The German page for Llama carries the sources with their retrieval dates.
Is Llama the right fit?
We go through your tasks and the route that carries them - and say plainly where it does not.
Arrange a conversation