General-Purpose AI Code of Practice

Summary

A voluntary code, prepared by independent experts, that providers of general-purpose AI models can sign to demonstrate compliance with Articles 53 and 55 of the AI Act. In its Transparency and Copyright chapters, which address all providers, signatories commit to keep model documentation up to date, give relevant information to downstream providers and, on request, the AI Office, and put in place a policy to comply with EU copyright law. The Safety and Security chapter, for providers of models with systemic risk only, sets out state-of-the-art practices for managing those risks, including a Safety and Security Framework, Model Reports and serious-incident reporting.

Safety topics

Status notes

Published on 10 July 2025, when the Commission received the final version, developed by 13 independent experts with input from over 1,000 stakeholders, and complemented by Commission guidelines on key concepts for general-purpose AI models. The AI Act rules it supports apply from 2 August 2025 and become enforceable by the AI Office one year later for new models and two years later for existing models. On 1 August 2025 the Commission, in an opinion, and the AI Board confirmed that the code is an adequate voluntary tool for providers to demonstrate compliance with the AI Act; adherence does not constitute conclusive evidence of compliance. The three chapters were authored separately, and the Transparency chapter includes a Model Documentation Form. Serious-incident reporting is Commitment 9 of the code, supported by a Commission reporting template of 4 November 2025. Signatories include Aleph Alpha, Amazon, Anthropic, Google, IBM, Microsoft, Mistral AI and OpenAI; xAI signed only the Safety and Security chapter and has to demonstrate compliance with the transparency and copyright obligations by other adequate means. Signatories have set up a Signatory Taskforce, chaired by the AI Office, for a coherent application of the code.

Sources

Open the official source https://digital-strategy.ec.europa.eu/en/policies/contents-code-gpai

Secondary source: https://digital-strategy.ec.europa.eu/en/library/commission-opinion-assessment-general-purpose-ai-code-practice

Local copy

Verification

History

1. 2026-10-18related: [{"slug": "eu-ai-act", "id": "EU-03"}, {"slug": "eu-guidelines-gpai-providers", → [{"slug": "eu-ai-act", "id": "EU-03"}, {"slug": "eu-guidelines-gpai-providers", — two-way link added: EU-02 link(s) to EU-102026-10-18-fix-related-links · Claude Code (claude-opus-5-5)
2. 2026-10-17notes: Published 10 July 2025 after a multi-stakeholder drafting process chaired by ind → Published on 10 July 2025, when the Commission received the final version, devel — rewritten from the stored files; added the experts and stakeholders and the application and enforcement dates from the press release; removed the count of 22 signatories (the English and Dutch versions of the list differ), the statement that Meta is not on the list and the 'GPAI CoP Appendixes' of Feb 20262026-10-17-recheck-europe-1 · Claude Code (claude-opus-5-5)
3. 2026-10-17safety_topics: ["general-purpose / frontier AI", "evaluation & testing", "incident reporting", → ["general-purpose / frontier AI", "transparency & labelling", "catastrophic risk — 'evaluation & testing' and 'accountability & liability' dropped because no stored text of this record shows them; 'incident reporting' kept for Commitment 9; 'catastrophic risk' kept for the chapter on systemic risks from the most advanced models2026-10-17-recheck-europe-1 · Claude Code (claude-opus-5-5)
4. 2026-10-17summary: A voluntary code that GPAI model providers can use to show compliance with AI Ac → A voluntary code, prepared by independent experts, that providers of general-pur — rewritten from the stored code page, Transparency chapter and Commission pages: the Copyright chapter and documentation commitments added; of the Safety and Security chapter's contents only the framework, model reports and serious-incident reporting are shown in stored files2026-10-17-recheck-europe-1 · Claude Code (claude-opus-5-5)

Related records

EU AI Act European Union · Partially in force

GPAI provider guidelines European Union · Published (non-binding, current)

GPAI training-data summary template European Union · Published (non-binding, current)

AI Act serious incident guidance (draft) European Union · Draft / proposed

European AI Office Decision European Union · In force

Citation

European Commission (AI Office), drafted by independent experts; assessed as adequate by the Commission and the AI Board. General-Purpose AI Code of Practice. 2025-07-10. https://digital-strategy.ec.europa.eu/en/policies/contents-code-gpai. Status: Published (non-binding, current); verified 2026-10-17 (level A). Local copy archive/EU-10/2026-10-11/official.html retrieved 2026-10-11 13:58:32 UTC, SHA-256 ebf9499d6044846160b2ab924c109e0b00d59cd0cd5cb516028ae00bba201fcc. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record EU-10.

Improve this record

Know of an amendment, a status change or an error? Report news or an error on GitHub; the form opens with this record filled in.