Comment by glub
8 hours ago
From my limited testing of just 2 hours, reasoning output of 5.1-max is at least 7x of 5-max, on the same project and comparable prompts.
It reasoned for ~2 minutes trying to figure out an appropriate directory name. I've never seen 5-max do that. Could be a misconfiguration though.
I did a bunch of Fable 5.1 xhigh review work on a bunch of critical components, ones with direct comparison from Fable 5 xhigh runs from two weeks ago. Token cost was 1.5-2x for each component.
Yeah, after some more testing, I think I'm going to pin it back to 5.
It feels like 5.1 is 5 that has higher reasoning threshold. I've been using fable as orchestrator anyway, so I see no reason to use 5.1.
I haven't dug deep but I burned through 30% of my weekly usage in a few hours which shocked me at first.