Anthropic says it met its AI R&D safety threshold
What it shows
A wall-mounted dial gauge with a curved scale and a needle; its green limit notch stands for a safety threshold Anthropic says it has met.
Its labels
- New: Threshold judged met
- Based on expected future models
- Human review of research decisions considered
The story
Anthropic says it met its AI R&D threshold, judged on expected future models · 12 Nov 2026
An illustration is never a picture of what happened. Drawn on 12 Nov 2026.