llmradar/eu-open-weight-models

EU-readiness of open-weight LLMs

Curated by LLM Radar — updated 2026-05-03 — 55 models.

A manually-reviewed dataset assessing open-weight Large Language Models (LLMs) on their suitability for EU deployment and commercial use. Each model is evaluated on licence, commercial use, training data, and origin, with quality / speed / price metrics from Artificial Analysis where available.

Primary use cases:

• Selecting open-weight models for self-hosted EU deployment
• Licensing and commercial-use due diligence
• Training classifiers on open-source LLM licensing and provenance
• Tracking the emergence of EU-origin and permissively-licensed models

Fields

column	description
name	Human-readable model name
vendor_name	Publisher / research lab
vendor_country	HQ country of the vendor
verdict	Overall verdict: ok, warn, or ko
verdict_text	One-line editorial verdict
licence	Licence summary (Apache-2.0, MIT, custom, etc.)
commercial_use	Commercial-use posture
training_data	Training-data disclosure / provenance
origin	Origin / publisher jurisdiction
tags	Comma-separated topic tags (permissive, commercial, euorigin)
quality_index	Artificial Analysis Quality Index (0–100, higher is better)
output_tokens_per_second	Median output speed (tokens/s)
blended_price_per_m	Blended input+output price per million tokens (USD)
context_tokens	Advertised context window
last_reviewed_at	ISO date of last manual review
llmradar_url	Canonical URL of the full model page
slug	Stable identifier on LLM Radar

Performance metrics (quality / speed / price) are sourced from Artificial Analysis under their public API and are attributed there; see their terms for reuse.

Methodology

Each entry is reviewed manually against public sources (provider docs, terms of service, data-protection agreements, sub-processor lists, regulator guidance, model cards, and vendor statements). Badges use a three-tier traffic-light scheme:

• green — meets the EU/GDPR/AI-Act criterion without material caveats
• amber — partial fit, conditional, or requires customer action to qualify
• red — does not meet the criterion, or requires data transfer outside the EEA

See the full methodology at https://llmradar.eu/methodology.

License & attribution

Released under CC BY 4.0. You may share and adapt the dataset for any purpose, including commercial use, provided you give appropriate credit and link back to the source:

Data from LLM Radar — https://llmradar.eu — licensed under CC BY 4.0.

Updates

This dataset is regenerated on a weekly cadence from the LLM Radar editorial database. Each row carries a last_reviewed_at date so you can filter for recency.

Citation

@misc{llmradar_eu-open-weight-models,
  title        = { EU-readiness of open-weight LLMs },
  author       = { {LLM Radar} },
  year         = { 2026 },
  howpublished = { \url{https://huggingface.co/datasets/llmradar/eu-open-weight-models} },
  note         = { CC BY 4.0 }
}

Report corrections

Found an inaccuracy? Submit a correction or exercise your right of reply at https://llmradar.eu/corrections or https://llmradar.eu/right-of-reply.