The European Commission has updated the signatory list of the General-Purpose AI Code of Practice, naming 22 providers that have signed the full code. The figures and names in this article come from the October 21, 2026 archived capture of the Commission's page, which was last updated on October 7, 2026; the live page could not be re-verified at publication time, so this report reflects that archived list, not a fresh confirmation of current membership. The Commission notes that names are added continuously as signatures are confirmed.
Who is on the list
The archived list names 22 full signatories: AI Studio Delta, Aleph Alpha, Almawave, Amazon, Anthropic, Black Forest Labs, Bria AI, Cohere, Domyn, Dweve, Fastweb, Google, IBM, LINAGORA, Microsoft, Mistral AI, Open Hippo, OpenAI, Pleias, Reflection AI, ServiceNow and WRITER.
In the archived list, xAI appears separately: it signed up only to the Safety and Security chapter. According to the Commission, that means xAI will have to demonstrate compliance with the AI Act's transparency and copyright obligations through alternative adequate means rather than through the code.
What signing the code means
The code of practice is a voluntary tool, prepared by independent experts in a multi-stakeholder process and published on July 10, 2025. Its Transparency and Copyright chapters give providers of general-purpose AI models a way to demonstrate their obligations under Article 53 of the AI Act; the Safety and Security chapter applies to the smaller group of providers whose models carry systemic risk under Article 55.
The Commission and the AI Board have confirmed that the code is an adequate voluntary tool for demonstrating compliance. Providers that sign can show they comply with the AI Act by adhering to the code, which the Commission says reduces their administrative burden and gives more legal certainty than proving compliance by other methods. Signatories have also formed a Signatory Taskforce, chaired by the EU AI Office, to coordinate how the code is applied.
Why it matters for builders
For startups choosing foundation models, the list is a practical reference: signatories have committed to the code's transparency documentation and copyright policy measures, which affects the information available downstream about training data and model behavior. Companies not on the archived list, including Meta, remain subject to the same AI Act obligations if they provide general-purpose models in the EU - they simply demonstrate compliance without the code's framework. The obligations themselves are not voluntary, and the list is worth re-checking as the Commission confirms further signatures.