chainoflack

[chainoflack @chainoflack]

@theo Nah bro, just get some rest and prepare for the heavy week that is about to come with 5.6 and it's daddy mythos

[chainoflack @chainoflack]

@amorriscode Wait. Is it already available?

[chainoflack @chainoflack]

@sunwukk990 @Lentils80 This is essentially untrue.

[chainoflack @chainoflack]

@Dodlanderx @mirochill I wouldn't call it close. Or you might, only in the same way you can call kimi close to gpt 5.5

[chainoflack @chainoflack]

@karlach_mary @mirochill This comment just made me realize how mythos can be so op that you might as well don't need an image generation capability lmao

[chainoflack @chainoflack]

@willmcgugan @OpenAI how ungrateful are you that you have to publish this instead of just telling them privately?

[chainoflack @chainoflack]

@johnhelmuth_ @argofowl they cant just fix it that easily, they have tried and only achieved mediocre results... they have to work way before the post training, and that takes time.

[chainoflack @chainoflack]

@ScarlettBeats @HarshithLucky3 sure buddy

[chainoflack @chainoflack]

@Marcia_Ong wrong, u will come back for mythos pretty soon

[chainoflack @chainoflack]

@jxnlco @petergyang Love ur honesty

[chainoflack @chainoflack]

@Aknotymous This is true. I do prefer oai models because they are superior in most things, but I cannot avoid coming back to claude every time because gpt takes me 10 inputs more for it to understand what I want vs opus, which instantly gets what I want to achieve

[chainoflack @chainoflack]

@Clipizo @thetreygoff Yes but with a tasteless criteria. Most people don't have jobs that are as repetitive to be worth writing so detailed instructions every time... This reminds me of those oai people who were proud of the gpts results when using their frontend skill mds lmao.

[chainoflack @chainoflack]

@thetreygoff They are improving on that. But I think the gap will be similar if not worse with the next models: I suspect gpt5.6 will be as capable as Mythos when a task is measurable, but Mythos will be so much easier to work with if you don't mind the price and are working with nuances

[chainoflack @chainoflack]

@teslascope :(

[chainoflack @chainoflack]

@atmoio you dont know how to read?

[chainoflack @chainoflack]

@testingcatalog im gonna prepare to spend all my money in this model. I just cant wait

[chainoflack @chainoflack]

@Conor_D_Dart @OpenAI i didnt know about this lol

[chainoflack @chainoflack]

@nicdunz Constantly

[chainoflack @chainoflack]

@kr0der @ajambrosino The more they try the worse it will get. There is something wrong with their pre training taste and they have to solve that first

[chainoflack @chainoflack]

@spencerkm @surface thank you!