# Uprelic: strom

> Uprelic's hosted decision API, run on GPUs in Paris. Answers noul, choice and score questions about text, JSON and up to 8 images, with a probability for every option and no generated text, on the same request shape as Jev. No weights; base model and size not disclosed.

- Page: https://systemonemodels.ai/uprelic/strom
- API: https://api.systemonemodels.ai/v1/models/uprelic/strom

## Facts

| | |
|---|---|
| Maker | Uprelic (https://systemonemodels.ai/uprelic) |
| Decides | choice, score, noul, classify, route |
| Architecture | strom |
| Context | 32K tokens |
| Licence | proprietary |
| Availability | Hosted API |
| Released | 2026-10-03 |
| Latest version | 2026.09 |

## Hosted API

- Provider: Uprelic (https://platform.uprelic.com)
- Docs: https://platform.uprelic.com/skill.md
- Input: $0.042/MTok · Output: Free

## Reported evaluation

Suite: JevBench public set (231 items), the maker's production run of Strom 1.0.7 on 1 Oct 2026 (Open-Jev harness format). Numbers are the publisher's own.

- Decision accuracy: 87.0%
- Calibration error (ECE): 0.033
- Median latency: 136 ms

## Model card

<!-- generated by scripts/seed_catalog.py; edit content/models/catalog.yaml -->

# Strom 1.0.7

Uprelic's hosted decision API, run on GPUs in Paris. Answers noul, choice and score questions about text, JSON and up to 8 images, with a probability for every option and no generated text, on the same request shape as Jev. No weights; base model and size not disclosed.

Uprelic describes Strom as a large vision-language model fine-tuned on business decisions in 14 domains, served on Scaleway GPUs in Paris with no third-party AI provider in between; it does not name the base model or size. Sign-up opened in October 2026 after a private beta; the public tier has no SLA during the beta. Up to 256 questions per request, 8 images and 32,768 tokens. The maker's own production run of JevBench's public items (a third-party benchmark) gives 87.0%, and it says two slices overlap its training data; on control suites rebuilt the Open-Jev way it says Jev is ahead. A dedicated EU deployment is offered under Strom Enterprise.

## What it decides

- **choice** — picks one option from a set
- **score** — places the input on an ordered scale
- **noul** — answers a yes/no question with one probability
- **classify** — assigns a category from a fixed taxonomy
- **route** — sends the input to one of several destinations

## At a glance

| | |
|---|---|
| Maker | Uprelic |
| Released | 2026-10-03 |
| License | proprietary |
| Reported accuracy | 87.0% |
| Reported latency | 136 ms median per answer from a client in Germany, network included |

## Hosted API

Served by **Uprelic** — $0.042/MTok input, $0/MTok output. [Get access](https://platform.uprelic.com) · [API docs](https://platform.uprelic.com/skill.md).

## Read more

- [Strom docs](https://platform.uprelic.com/skill.md)
- [Launch post](https://platform.uprelic.com/blog/strom-eu-alternative-decision-model-to-jev)
- [Python SDK](https://pypi.org/project/uprelic/)

---

*This page was opened by System One for Uprelic, who can claim the organisation and take it over at any time.*

---

From System One Models — https://systemonemodels.ai/ · every System One model: https://systemonemodels.ai/system-one-models