We value your privacy

We use cookies to enhance your browsing experience, serve personalized content and analyze our traffic. By clicking “Accept all” you consent to our use of cookies. Cookie Policy

More options
Skip to content
FacthoryFacthory
FacthoryFacthory
    • Workspace
    • Agent
    • Knowledge
    • Analytics
    • Office apps
    • Desktop
    All features
    What’s new
    Download
  • By work

    • Organizational Knowledge Risk & Gap Intelligence
    • Scientific Research & Experimentation
    • Autonomous Analytics & Insight Discovery
    • Workforce Readiness & Adaptive Learning
    • SOP Governance & Operational Standardization
    • Process Discovery & Continuous Optimization
    • All solutions

    By industry

    • Manufacturing
    • Automotive
    • Construction & Infrastructure
    • Real Estate
    • Public Sector
    • Life Sciences
    • All industries
  • Facthory for

    • Regulated industries
    • On-premises
    • Developers

    Discover

    • What’s new
    • Blog
    • Case studies
    • Webinars
    • Models

    Learn

    • Documentation
    • Help centre
  • Pricing
  • Download
Sign inGet started
Models

Know a model before you run it

Facthory scans open models for vulnerabilities, scores each against the EU AI Act, and measures what it can do. One page per model, every fact with its source.

Browse the catalogModels radarHow we scoreAsk about a model

Three scores for every model

Vulnerability scanning

What loading the model can do to your machines, and whether its files are what they claim to be.

  • Weight formats: pickle files can run code when they load, safetensors cannot
  • The hubs’ own malware and pickle scans of every file
  • Provenance: publisher, release, checksums and signatures
  • Licence terms, and known issues published for the model

EU AI Act scoring

Every model scored against the EU AI Act’s obligations for general-purpose AI models, article by article.

  • Article 51: whether the model has systemic risk
  • Article 53 with Annexes XI and XII: documentation, copyright policy, training summary
  • Article 55: evaluations, adversarial testing, incidents and cybersecurity
  • Articles 4, 26 and 50: what is left for you as the deployer

Capability scoring

What the model is good at, from published benchmarks, and what it needs to run.

  • Coding, maths, reasoning, knowledge and conversation, scored by category
  • Languages, German included
  • Memory and GPUs at each quantization, for running it on your own hardware
  • Each benchmark with its source and version
EU AI Act

The articles we score against

The EU AI Act (Regulation (EU) 2024/1689) sets obligations for providers of general-purpose AI models since 2 August 2025, and for those who deploy AI systems. Each model page shows, article by article, what its provider has published and what remains yours to do.

ArticleWhat the Act asksWhat we checkBinds
Art. 51

Systemic risk

A general-purpose AI model has systemic risk when it has high-impact capabilities, presumed above 10²⁵ floating-point operations of training compute, or when the Commission designates it.

Published training compute, parameters and data size, and the AI Office’s designations, to tell which obligations apply.

Provider
Art. 53(1)(a), Annex XI

Technical documentation

The provider keeps technical documentation of the model, its training and testing, and the results of its evaluation, for the AI Office and national authorities.

Whether the provider states that it keeps it, and which of Annex XI’s items its public documentation covers.

Provider
Art. 53(1)(b), Annex XII

Information for downstream providers

The provider gives those who build on the model what they need to understand its capabilities and limits and to meet their own obligations.

The model card and documentation against Annex XII: intended and acceptable uses, architecture, parameters, modalities, context length, licence and a description of the training data.

Provider
Art. 53(1)(c)

Copyright policy

The provider puts in place a policy to comply with Union copyright law, including rights holders’ reservations against text and data mining.

A published copyright policy, and how it honours opt-outs from text and data mining.

Provider
Art. 53(1)(d)

Summary of training content

The provider makes public a sufficiently detailed summary of the content used for training, following the AI Office’s template.

The published summary, and whether it follows the template.

Provider
Art. 53(2)

Open-source exemption

Models released under a free and open-source licence with their weights public are exempt from points (a) and (b), unless they have systemic risk.

The licence’s terms and whether the weights, architecture and usage information are public.

Provider
Art. 54

Authorised representative

A provider established outside the Union appoints an authorised representative in the Union before placing its model on the Union market.

Where the provider is established, and the representative it names.

Provider
Art. 55

Obligations for systemic-risk models

The provider evaluates the model, adversarial testing included, assesses and mitigates systemic risks, reports serious incidents, and protects the model’s cybersecurity.

Published evaluations and red-teaming results, the provider’s risk and incident reporting, and its security measures.

Provider, systemic risk
Art. 56

Code of practice

Providers may show compliance through codes of practice, such as the General-Purpose AI Code of Practice.

Whether the provider has signed the Code of Practice, and which chapters.

Provider
Art. 50

Transparency of AI output

People must know when they interact with an AI system, and synthetic audio, images, video and text are marked as generated.

Whether the provider documents how the model’s output can be marked or detected, which a deployer needs for its own duties.

Provider and deployer
Art. 4

AI literacy

Providers and deployers take measures so that their staff who operate AI systems have sufficient AI literacy.

Not a property of the model: listed so a deployer sees the duty beside the model it chooses.

Deployer
Art. 26

Deployers of high-risk systems

A deployer of a high-risk AI system uses it according to its instructions, assigns human oversight, monitors it and keeps its logs.

Whether the provider’s documentation gives the instructions, limits and logging a deployer needs when the model serves a high-risk use (Annex III).

Deployer

When the obligations apply, as the Regulation sets them (Art. 113)

Amendments can move these dates; each model page cites the text in force when it was scored.

2 February 2025

Prohibited practices (Art. 5) and AI literacy (Art. 4).

2 August 2025

Obligations for providers of general-purpose AI models (Chapter V).

2 August 2026

Most other obligations, high-risk systems in Annex III among them.

2 August 2027

High-risk systems under Annex I, and models placed on the market before 2 August 2025 (Art. 111(3)).

Facthory Verified

A model earns the badge when its safety, EU AI Act and capability scores clear published thresholds and a person has reviewed the evidence behind them.

  • Thresholds published with the scores
  • Re-checked when the model or its sources change
  • Withdrawn when a finding is open

Scores describe published evidence about a model. They are not legal advice, and a missing document is not proof of a breach. How we score

A workspace for what your organization knows, with an agent that works on it.

Product

  • All features
  • Workspace
  • Agent
  • Knowledge
  • Analytics
  • Office apps
  • Desktop
  • Download
  • Pricing

Solutions

  • Organizational Knowledge Risk & Gap Intelligence
  • Scientific Research & Experimentation
  • Autonomous Analytics & Insight Discovery
  • Workforce Readiness & Adaptive Learning
  • SOP Governance & Operational Standardization
  • Process Discovery & Continuous Optimization
  • All solutions
  • All industries

Resources

  • Blog
  • Case studies
  • Webinars
  • Documentation
  • Help centre
  • Models
  • ROI calculator

Company

  • Who we are
  • Careers
  • Press hub
  • Brand kit
  • Philosophy
  • Contact us

Join our newsletter

Product news from Facthory, by email.

​

© 2026 Facthory. All rights reserved.
  • Privacy policy
  • Cookie policy
  • Terms and conditions
  • Imprint
  • Cookie settings