[seqfan] Problem

[Tomasz Ordowski]

Dear SeqFans,

I have a question:

 Are there numbers n such that lambda(n) < phi(n) and
b^{lambda(n)/2} == +-1 (mod n) for every b coprime to n ?

Best regards,

Thomas Ordowski
___________________________________________
Amiram Eldar did not find any such number n < 10^5.