EU AI Act GPAI Code of Practice
General-purpose AI obligations enforced from Aug 2025
The AI Act's rules for general-purpose AI (GPAI) models entered into force in August 2025, with a voluntary Code of Practice (co-signed by OpenAI, Anthropic, Google, Mistral, and others; not signed by Meta) covering transparency, copyright, and systemic-risk (models trained above 10^25 FLOPs). It is the world's first binding horizontal AI law and directly shapes how frontier labs document training data and safety. High-risk system obligations phase in through 2026-2027.
GPAI rules live
Aug 2025
Systemic threshold
>10^25 training FLOPs
Enforcer
European AI Office
How it fits the stack
EU AI Act GPAI Code of Practice with what it depends on (above) and what it feeds (below). The figure renders as a crawlable diagram and upgrades to an interactive 3D graph as it scrolls into view.
EU AI Act GPAI Code of Practice in the AI stack. EU AI Act GPAI Code of Practice with its immediate upstream dependencies (top) and downstream dependents (bottom) in the AI value chain. Hover a node in 3D, or read the full relationships below.
Graph data (text) — 3 entities, 2 relationships
- EU AI Act GPAI Code of Practice —depends on→ Content Licensing & Copyright Regime
- EU AI Act GPAI Code of Practice —depends on→ EU AI Act