Summarisation

Condensing long documents, minutes and reports without inventing content.

The router maps a task to candidate models, then applies your policy: region first, then price, then the model with the better documented result.

Prefer a context window that holds the whole document in one pass and an EU-operated endpoint for internal material; summarisation is where hallucination is most visible.

{"model":"llmeu-auto","messages":[{"role":"user","content":"…"}],"llmeu":{"task":"summarization"}}

Recommended

Mistral Small 4 (119B-A6B)

Mistral AI
Not hosted — mock backend Open weights European provider

Mistral Small 4 is an Apache-2.0 mixture-of-experts model with 119B total parameters and 6.5B activated per token. It accepts text…

The European model catalogue