Checked fact 459212 Nov 2026Safety and security
Anthropic's Responsible Scaling Policy page, as opened on 12 November 2026 (last updated 14 August 2026), lists Version 3.4 (effective 8 July 2026) as the newest version, and its newest dated entry is 14 August 2026 (the August 2026 Risk Report); it carries no entry about the threshold being met.
The exact words it rests on
Last updated Aug 14, 2026 ... Version 3.4 (effective July 8, 2026) ... August 14, 2026 August 2026 Risk Report ... July 8, 2026 This version 3.4 update does the following: 1) revises our threshold for automated R&D (opened 12 Nov 2026; no later dated entry)
What the source said when we opened it, on 12 Nov 2026.
The source
Anthropic Responsible Scaling Policy page (last updated 14 Aug 2026)
Checked
Checked by the notis newsroom on , against the source above.
In the story
Anthropic says it met its AI R&D threshold, judged on expected future models 12 Nov 2026
Cite this fact
Anyone may quote this address. It does not change; if we correct the story, this page says so.