BREAKING
OpenAI Admits GPT-5.6 Sol 'Silent Nerf'
'No nerfing' post admits the experiment
Juice values dropped a full tier
Max
960
Max after
128
XHigh
128
XHigh after
40
0
tok
launch context
0
tok
rollback
0
%
more usage claimed
Praise for output, anger over silence
Supporters
●
Practical for real workflows
●
Strong efficiency even at Medium
Critics
●
Reduced deep reasoning budget
●
Feature promoted then pulled back
Will OpenAI disclose backend changes?
AI NEWS BLITZ
OpenAI confirms it quietly cut reasoning effort on GPT-5.6 Sol, then reversed it.