WHAT NEW IN X10.31 SAT

AI collections from X · 22 collections · 63,170 posts

jevIMAGE

Jev takes over yes/no decisions in the agent loop, Opus only plans and writes code

Judgments like which model tier to use, whether a command is safe, and whether the task is done go to Jev at $0.042 per million tokens, while Opus is kept for planning and writing code.

CyrilXBT
@cyrilXBTArchived 17 days agoOriginally posted 2026.10.13
Jev takes over yes/no decisions in the agent loop, Opus only plans and writes code
CyrilXBT @cyrilXBTOriginally posted 2026.10.13
Open original on X

most people will read this and think it’s about a cheaper model it’s not. it’s about which questions deserve a model at all → “which tier takes this?” doesn’t need Opus → “is this command safe?” doesn’t need Opus → “is it done?” doesn’t need Opus that’s most of your agent loop. Opus plans and writes the code. Jev answers everything else for $0.042 per million tokens.

This post quotes another post: x.com/1373665408193036293/status/210869352490011…

x.com/cyrilxbt/status/2109898184432709938Archived 2026.10.14 10:10