As the U.S. and China continue to speed ahead with new AI developments, spearheaded by OpenAI and DeepSeek, Europe has somewhat fallen into the background of the global AI race.
In an attempt to change this, a European alliance has announced the development of OpenEuroLLM—a European group of large language foundation models.
The aim is to become a competitive alternative to OpenAI and DeepSeek while remaining grounded in the EU’s stringent safety and privacy guidelines.
However, with OpenAI’s ChatGPT and DeepSeek’s R1 leading the AI race with millions of users, how does OpenEuroLLM currently compare?
Developed by European AI allies, OpenEuroLLM is designed as an open‐source alternative that strengthens Europe’s independent AI capabilities.
The mission is to reduce reliance on U.S. and Chinese models by fostering collaborative, home‐grown AI innovation in the continent.
OpenEuroLLM has said it would incorporate AI into European values, focusing on transparency, openness, and democracy.
Developed by OpenAI, ChatGPT is part of the AI firm’s mission to “ensure that artificial general intelligence—AI systems that are generally smarter than humans—benefits all of humanity.”
It is offered in both free and subscription-based versions, with continuous refinements driven by extensive usage feedback.
ChatGPT is designed to be a widely accessible, safe, and useful tool that helps people solve problems and streamline everyday tasks.
Thousands of companies use OpenAI’s chatbot, which is strong in generating content and handling general-purpose inquiries.
Created by a Chinese research team, DeepSeek emphasizes cost efficiency and rapid deployment.
It is offered free and is built to perform well on technical, logic‐oriented tasks.
However, it has also drawn attention to issues such as strict censorship on politically sensitive topics and data privacy concerns, given that user data is stored on servers in China.
With European data protection as a core design principle, OpenEuroLLM will be expected to adhere to the sweeping rules being enforced in the continent.
The EU’s AI Act, which came into force on August 1, requires transparency for all AI systems that pose potential risks.
All AI systems deemed high-risk will need to comply with specific legal requirements, while those deemed unacceptable will be banned.
OpenEuroLLM, which focuses on transparency and ethical development, will likely play a pivotal role in the EU’s heavily regulated AI landscape.
ChatGPT adheres to Western data protection standards and is subject to ongoing debates over bias and data handling.
OpenAI’s closed-source nature means that many of its internal processes remain proprietary, sometimes limiting transparency compared to open-source models.
Despite its technical prowess, DeepSeek has faced regulatory scrutiny in Italy and Ireland over its data storage practices.
The fact that user data may be stored on servers in China, combined with the model’s built-in censorship mechanisms, has raised concerns about transparency and user rights.
While detailed technical documentation on the project is still emerging, OpenEuroLLM is built with a focus on transparency and open collaboration.
The EU AI Act states that any AI developers acquiring data for AI training from the web must ensure they have first received consent.
This means that OpenEuroLLM’s architecture is unlikely to scrape as much data from the web as ChatGPT and DeepSeek.
ChatGPT uses a traditional transformer-based architecture that processes queries at the full capacity of the model.
This type of architecture demands vast computational resources, meaning it comes with higher training and operational costs.
According to 2023 estimates from research firm SemiAnalysis , ChatGPT cost OpenAI around $700,000 daily to run, but in 2025 a single query can now cost $1,000.
DeepSeek employs a Mixture-of-Experts (MoE) architecture, which means that for each query, only a subset of its 671 billion parameters is activated.
This selective processing significantly reduces training and operational costs and allows it to excel in technical tasks and logical reasoning.
“Think of it like assembling the Avengers for every input — only the best-suited heroes (experts) get called in,” Harshada Jivane, machine learning engineer, explained in a blog post.
The new European AI project currently has a budget of €52 million, along with a compute commitment of a possibly higher value, Peter Sarlin, the co-founder of Silo AI who is backing the project, told TNW .
In addition to being backed by the European Commission, OpenEuroLLM has partnered with over 20 research institutions and leading AI startups.
The project also receives funding from STEP, an EU scheme to boost investment in innovative technologies.
Over the years, OpenAI has attracted substantial investments from major backers, including Microsoft, SoftBank, and Oracle.
According to the Wall Street Journal , the U.S. AI firm is in talks to raise a new round of funding to double its valuation to $340 billion.
These billions of dollars of investments underscore the heavy capital requirements behind building and running advanced AI models like ChatGPT.
In contrast to the enormous sums raised by OpenAI, DeepSeek is reportedly developed on a much leaner budget.
DeepSeek’s development is fully funded by High-Flyer , a quantitative hedge fund in China. According to public disclosures, its flagship model, R1, was trained with an investment of roughly $9 million.
However, some technology experts have claimed the actual price of development may have been a lot higher.