EU AI sovereignty tracker

Which AI can you actually ship in Europe?

Independent evaluation of API providers and open-weight models against European hosting, GDPR and licensing requirements. Updated monthly, sourced from official documentation.

Answer a few questions, get a ranked shortlist in 30 seconds. Try the advisor →
Use this tab if you self-host model weights. The question is: is this model legally and operationally safe to deploy on your infra? We evaluate licence, restrictions, training data transparency and supply-chain risks.
Showing 55 of 55 models
ModelLicenceCommercial useOriginCompare
Talkie-1930-13B Base
Apache 2.0UnrestrictedUS (research)
IBM Granite 4.1 30B
Performance
Apache 2.0UnrestrictedUSA
IBM Granite 4.1 8B
Performance
Apache 2.0UnrestrictedUSA
Ling-2.6 1T
Performance
MITUnrestrictedChina
Ling-2.6 Flash
Performance
MITUnrestrictedChina
Laguna XS.2
Apache 2.0UnrestrictedUSA
Nemotron-3 Nano Omni 30B-A3B Reasoning
Performance
NVIDIA Open ModelPermitted (with attribution)USA
Mistral Medium 3.5
Performance
Modified MITPermitted (revenue cap)EU (France)
Talkie-1930-13B-IT
Apache 2.0UnrestrictedUS (research)
Hy3-preview
Performance
Tencent Hy CommunityEU territory excludedChina
MiMo-V2.5
Performance
MITUnrestrictedChina
MiMo-V2.5-Pro
Performance
MITUnrestrictedChina
Kimi-K2.6
Performance
Modified MITAttribution at scaleChina
Qwen3.6-27B
Performance
Apache 2.0UnrestrictedChina
DeepSeek-V4-Flash
Performance
MITUnrestrictedChina
DeepSeek-V4-Pro
Performance
MITUnrestrictedChina
Mistral 7B Instruct v0.2
Performance
Apache 2.0UnrestrictedEU (France)
GPT-OSS 20B
Performance
Apache 2.0UnrestrictedUnited States
Qwen3-8B
Performance
Apache 2.0UnrestrictedChina (Hangzhou)
Llama 3.1 8B Instruct
Performance
Llama 3.1 Community LicenceRestricted (MAU cap + AUP)United States
Gemma 4 E4B Instruct
Performance
Apache 2.0UnrestrictedUnited States
Gemma 4 26B A4B Instruct
Performance
Apache 2.0UnrestrictedUnited States
Gemma 4 31B Instruct
Performance
Apache 2.0UnrestrictedUnited States
GLM-5.1
Performance
MITUnrestrictedChina (Beijing)
Qwen3.6-35B-A3B
Performance
Apache 2.0UnrestrictedChina (Hangzhou)
MiniMax-M2.7
Performance
MiniMax Non-Commercial LicenseNon-commercial onlyChina (Shanghai)
Kimi K2 Instruct
Performance
Modified MITYesChina
Grok-2
Performance
Grok 2 CommunityYes (w/ AUP)USA
SmolLM3 3B
Apache 2.0YesUSA
MiniMax M2
Performance
Modified MITYesChina
Falcon H1 34B
Falcon-LLMYesUAE
Gemma 4
Performance
Apache 2.0YesUSA
GLM-4.5
Performance
MITYesChina
DBRX Instruct
Performance
Databricks OpenWith capsUSA
OLMo 2 32B
Performance
Apache 2.0YesUSA
Jamba 1.5 Large
Performance
Jamba OpenUnder $50M rev.Israel
Llama 3.1 Nemotron 70B
Performance
Llama communityWith capsUSA
DeepSeek V3
Performance
MITYesChina
DeepSeek V3.2
Performance
MITYesChina
IBM Granite 3
Performance
Apache 2.0YesUSA
Command R+
Performance
CC-BY-NC 4.0API onlyCanada
QwQ-32B
Performance
Apache 2.0YesChina
Qwen 3.5
Performance
Apache 2.0YesChina
DeepSeek R1
Performance
MITYesChina
Qwen 2.5
Performance
Apache 2.0YesChina
Llama 4 Scout
Performance
Llama communityWith capsUSA
Llama 4 Maverick
Performance
Llama communityWith capsUSA
Phi-4
Performance
MITYesUSA
Mistral Small 4
Performance
Apache 2.0YesEU
GPT-OSS 120b
Performance
Apache 2.0YesUSA
Llama 3.1 405B
Performance
Llama communityWith capsUSA
Llama 3.3 70B
Performance
Llama communityWith capsUSA
Codestral 22B
MNPL (non-prod)Paid licence req.EU
Mistral Large 2
Performance
MRL (research)Paid licence req.EU
Mistral Small
Performance
Apache 2.0YesEU
Home·Methodology·Advisor·Changelog·About·Press·Privacy·