Positive users praise GPT-5.6 Luna models for delivering better value and performance than pricier Terra or Sol variants, while negative users criticize the added complexity of choosing among many model and effort combinations.
Based on 41 visible X reactions from 257 accounts.
Ask a question below.
Published answers will appear here.
@rasbt I love Luna, I hope it stays like this so I can get through my whole week without running out of my limit on the Plus subscription.
@rasbt All of these stupid graphs need a time cost. Otherwise, they look really dumb.
@rasbt This sucks, with 5.5 everyone used xhigh on fast without worrying, now this nonsense.
@rasbt Excellent analysis!
The analysis recommends Luna over intermediate tiers like Terra.
@rasbt What a goddamn mess
@rasbt Nice graph
For agentic coding, one can say: - Unless you need Terra Ultra perf, it's always better to use a Luna model with higher effort setting (same or better performance but cheaper). - Forget everything below Sol High, use Luna with higher effort settings here - Forget Sol Extra High, use Terra Ultra here - The extra cost of Sol Ultra is probably not worth it over Max
Whomever @OpenAI decided that Ultra was higher than Max needs to take an optimization class. https://twitter.com/rasbt/status/2075573860796436626
Meanwhile, the open models keep on keeping pace. It is as if there's no wall to incremental improvement. https://x.com/michaeljburry/status/2075565375274529099
The charts don't lie! (I suspect) https://twitter.com/rasbt/status/2075573860796436626
@DhruvBatra_ @OpenAI It's the best best... as we used to say as kids.
It’s simple https://twitter.com/rasbt/status/2075573860796436626
Positive users praise GPT-5.6 Luna models for delivering better value and performance than pricier Terra or Sol variants, while negative users criticize the added complexity of choosing among many model and effort combinations.
Based on 41 visible X reactions from 257 accounts.
Ask a question below.
Published answers will appear here.
@rasbt Nice graph