Loading the catalogue…
Loading the catalogue…
Swedish cloud provider offering shared inference for open-weight models (Think) over an OpenAI-compatible API, from its own data centres in Sweden.
Five dimensions
Bands only. The scores and Stav's notes per dimension are read in the workspace.
Stav's narrative, every risk item with its source, the safeguards that offset them and the privacy-policy issues found: the part of the page a procurement review runs on.
Stav's narrative · 10 risk items · 6 safeguards · 5 privacy-policy issues · DPA, terms, privacy policy — sign in to read them.
Traffic through Stav is covered by Stav's customer DPA, which lists Stav's sub-processors.
This level applies to traffic routed through Stav.
evroc AB is a Swedish aktiebolag (559398-0930) serving the evroc Think inference product from EU data centres in Stockholm, Paris and Frankfurt, with no US parent or CLOUD Act exposure — strong L1 and l2.anti_coercion credentials. The binding L2 gate falls on independently audited security certification: no ISO 27001, SOC 2 Type II or equivalent certificate has been published, which is consistent with the platform having only reached general availability in mid-2025 and Think remaining in public beta. To advance to L3, evroc would additionally need to resolve the partial ownership picture (Giant Ventures UK minority stake) at the primary register level and publish evidence that privileged access is confined to EU/EEA personnel.
One criterion from L2: Independently audited security certification.
Stav's own assessment against criterion registry 4.1. Indicative, not an official recognition, which only a national authority can confer. Criteria follow the proposed Cloud and AI Development Act and are expected to change. “Union” reads as EU/EEA at every rung. CADA itself still says “in the Union”: until the Act is incorporated into the EEA Agreement, a Norwegian, Icelandic or Liechtenstein route does not evidence formal Union assurance for an EU public-sector buyer. UK and Swiss providers are capped at L1.
Verified 25 Sept 2026About the ladder →| Model | Input / M | Output / M | TTFT | Tok/s |
|---|---|---|---|---|
| glm-5-2 | €1.25 | €5.00 | 0.12 s | 307 |
| mistral-medium-3.5-128b | €1.50 | €6.00 | 0.38 s | 57 |
| Kimi K2.6 | €1.25 | €5.00 | 0.10 s | 189 |
| Gemma 4 26B A4B It | €0.13 | €0.50 | 0.10 s | 259 |
| Qwen3 Embedding 8B | €0.10 | €0.10 | 0.08 s | 0 |
| Qwen3-Reranker-4B | €0.05 | — | — | — |
| kb-whisper-large | — | — | 0.17 s | 0 |
| gpt-oss-120b | €0.20 | €0.80 | 0.09 s | 306 |
| multilingual-e5-large-instruct | €0.10 | €0.10 | 0.06 s | 0 |
| qwen3.8-27b | €0.75 | €3.00 | 0.11 s | 348 |