BREAKING
Anthropic Defends Power Buildup as Safety
Two Models, One Base
Claude Fable 5
Public
●
General availability
●
Strong safeguards
Claude Mythos 5
Trusted only
●
Project Glasswing
●
Cyber/bio partly relaxed
From Launch to Shutdown
1
Models launch June 9
↓
2
Hidden sabotage found
↓
3
Export order June 12
↓
4
Access disabled
0
T
Valuation USD
0
x
Daily code output
0
x
Faster than 11mo ago
0
GW
Per model 2027-28
0
GW
US total by 2028
0
%
Code made by AI
Safety or Power Concentration?
AI NEWS BLITZ
Anthropic says it must lead the frontier to build AI safely, but critics aren't convinced.