Comparison between Kimi K3.1 Pro (Max) vs Opus 5.5 (All effort values). AA numbers.
Seems like Kimi K 3.1 Pro Max is very comparable to Opus 5.5 High across the boards in numbers.
Will be posting more results soon of actual real world use case.
@Da7_Tech My thoughts too. It's amazing model for creative writing, unpredictable as well and that unpredictability which is good for writing isn't good coding where you need predictibilty.
I still sometimes lean towards using GPT 5.6 Sol (Whenever I use gpt models) rather than 6.1 Sol. You just feel the newer Sol models are much smaller ones. The newer ones just doens't have that 'instinct' and thoroughness.
@alexandr_wang Uhh... more like Muse is free so younger audiences also can use it, and Grok bot is paid so its more popular among older audiences.
I am not sure how that's a win for the Muse, mate...
Everyone has been saying Sonnet 5.5 is routing to Fable 5.5.
Did it cook or nah?
> Prompt: Lizard riding a bicycle with a fishing rod, 3js, web animation.
@LuminaBench > I doubt OpenAI would release something trash now and they have had a good amount of time to train it since their frontier RL training resumed
Remember how Sol 6 was, 6.1 ain't much better but still. They also have time from 5.6 to 6.
I dont have faith but lets see.
Everyone has been saying Sonnet 5.5 is routing to Fable 5.5.
Did it cook or nah?
> Prompt: Lizard riding a bicycle with a fishing rod, 3js, web animation.