Search pages, courses, and articles
Article 55 binds providers of general-purpose AI models with systemic risk: model evaluations, adversarial testing, serious incident reporting, cybersecurity. The 10^25 FLOP compute threshold triggers classification. Full Commission enforcement from 2 August 2026.
Article 55 sits in the GPAI provider lane, OpenAI, Anthropic, Google, Mistral, Meta. Most readers of this site are deployers, not GPAI providers, but understanding Article 55 matters because it shapes what GPAI providers can offer, what documentation they must hand down to downstream deployers via Article 53(1)(b), and what serious-incident reporting cascades when something goes wrong. The voluntary GPAI Code of Practice covers most of the operational detail. 2 August 2026 is when full Commission enforcement starts and the Omnibus did not change that.
Article 55 of the EU AI Act binds providers of general-purpose AI models with systemic risk: model evaluations, adversarial testing, serious incident reporting under Article 55(1)(c), and cybersecurity. The 10^25 FLOP compute threshold triggers classification. Full Commission enforcement starts 2 August 2026. The Omnibus deal did not change this regime.
Read articleA practical guide to AI incident response under the EU AI Act. 6-step response plan, Article 26(5) reporting obligations, whistleblower protections, and real incident examples.
Read articleOn 7 May 2026 the EU Council and Parliament reached a provisional agreement to delay the high-risk AI deadlines. Annex III standalone systems now apply from 2 December 2027, Annex I embedded systems from 2 August 2028. Article 4 literacy and Article 5 prohibited practices are unchanged and still live.
Read articleEU AI Act penalty tiers explained. Article 99 fines up to EUR 35 million or 7% of turnover, enforcement mechanisms, civil liability, and how to avoid penalties. Series finale.
Read articleMore topics
5-minute scored assessment across People, Tools, Compliance, and Infrastructure. See where you sit on this topic in operational terms.