system1models
Get API key

EU tier · GDPR · EU AI Act

Decision models.
Hosted in the EU.

For teams with the strictest data-protection requirements: EU-only inference on contracted EU operators, a Data Processing Agreement at signup, and no storage of prompts or answers.

Start free — 100M tokens Read the DPA
Request → responseIllustrative samples
REQUEST / choice

state: Mia owns a red bicycle.

question: Which color is the bicycle?

ANSWER / probabilities
red0.91
blue0.09
REQUEST / choice

state: a business customer asks to update an invoice address.

question: Which support queue fits?

ANSWER / probabilities
billing0.87
technical0.09
account0.04
REQUEST / noul

state: a viewer deletes a workspace.

question: Is this action allowed?

ANSWER / P(true)
true0.04
Illustrative contract shapes, not measured output — all samples and copyable code are in the quickstart.

Measured p95 latency

s1-fast244 ms
s1-pro644 ms

Local model-server time from a test with two concurrent requests and 1,000 mixed requests; API gateway and network time are excluded. This is not a service guarantee. How we measured.

Certifications held by our hosting providers.

Data flow

Where your data goes

Every EU request stays in Finland from the first byte to the last.

  1. Your appClient
  2. TLS 1.3
  3. API gatewayHelsinki, FI
  4. WireGuard
  5. GPU inferenceFinland

Answer returns over the same encrypted path

Not stored after the responsenot loggednot used for training

Contracts

What you get in writing

DPA

Data Processing Agreement under Art. 28 GDPR, accepted at signup.

Read the DPA →

EU AI Act

Aligned with the EU AI Act

System1 Models hosts open-weight general-purpose models for narrow decision tasks. We publish model cards and benchmark results, label our models clearly, and document intended use. Using the API for high-risk purposes under the AI Act requires your own assessment.

See model cards →

Independent benchmarks

Close to the Jev reference. For up to 40% less.

We publish results from several sources, including the ones where we trail. JevBench is run by our founder — we label it and never use it alone.

JevBench 1.5.1run by our founder

Ahead

of reference

s1-pro 73.2 vs 72.1 composite

Published model evidence · five equally presented sources · checked 30 September 2026
Model profileDecision Index0.2.1Chance-corrected index / 100Independent communityWorkflow Evalsdatasets 2026-09-28 · code 0ac3b8aModel-reference agreementTypeSafe · reference vendorJev Rerank Bench2026-09-25Dataset-macro nDCG@10 / 1Independent authorJevBench1.5.1Composite score / 100Same founder as System1ImageJevBench0.1.4Composite score / 100Same founder as System1
s1-fastPlumb-4BNot measured¹Not measured¹Not measured¹71.56Not measured¹
s1-proWinnow-12B Q850.02Not measured¹0.63473.2343.56
TypeSafe referenceJev 1.13.057.9167.80.67072.13Not measured¹

¹ Not measured means no comparable published measurement was verified for this model; it is not a zero. Results concern the tested model implementations, not the System1 service. ImageJevBench uses a separate image path; image input for customers is coming soon.

Conflict of interest: JevBench and ImageJevBench are run by the founder of System1 Models. Workflow Evals is published by TypeSafe AI, the vendor of the reference model. It measures model-reference agreement, not human-labelled accuracy. No overall score is published while comparable coverage is incomplete. Method and sources →

All benchmark sources and methods →

EU pricing

Clear prices for the EU tier.

Per 1 million input tokens. Output tokens are free. New accounts get 100M free input tokens.

ModelUSDEUR
s1-fast
Plumb-4B · Text		$0.034€0.030
s1-pro
Winnow-12B Q8 · Text		$0.040€0.035

Excluding applicable taxes. Top-ups: Germany today; EU businesses and more countries soon. All prices and calculator →