WKups

[WKups @kups_w]

@lydiahallie looking at https://t.co/QZZQXkEi4V that is the sequel of the sonnet 5 failure... :/

[WKups @kups_w]

@thsottiaux Why Astra is like a sine curve? Once is like AGI, another time...

[WKups @kups_w]

@lydiahallie Good job, Lydia :)

[WKups @kups_w]

@jvkashwanth @LDT0545 not really IMHO. at least cheaper

[WKups @kups_w]

@LDT0545 opus. GPT-6 sol is a slope...

[WKups @kups_w]

@ClaudeDevs Guys, what is the point of that... I want to fix sql injection IN OUR APP, and the model drops...

[WKups @kups_w]

@thsottiaux What about tokens generated? 5.6 luna was told to be as cheap as 5.4-nano, but at the same time it produced 30% more tokens

[WKups @kups_w]

@prywatnik bedzie zabawnie jak jednak nie wyjdzie :)

[WKups @kups_w]

@claudeai What is opus limit?

[WKups @kups_w]

@thsottiaux Since last friday, despite multiple restarts, the app hang on updating - 0% progress. What did you guys put there...

[WKups @kups_w]

@SolipsRenatus @thsottiaux @stemonteduro So true...

[WKups @kups_w]

@thsottiaux @stemonteduro credits are not maintanable - these are for such a small volume of work it does not make sense. Just calculate.

[WKups @kups_w]

@thsottiaux @romainhuet @sama I literally cannot wait for that, same as codex which is updating since yesterday.
Guys, quality first.

[WKups @kups_w]

@thsottiaux so far my codex is updating for the last 1h and it is still on 40%.

[WKups @kups_w]

@pradeepXkapoor

[WKups @kups_w]

@pradeepXkapoor the problem is with fable 5.1, not the usage in general. Read first what others report.

[WKups @kups_w]

@lydiahallie how much it adds to the cost? Did you measure it? I am afraid of keeping the context. Yup - that is something what i was looking for, but when I see how quickly tokens drains even on x20, I need to weighten capabilities. Either comfort or work done.

[WKups @kups_w]

@buildwithrajath final cuntdown.. I have noticed today in codex they added a popup "you are close to the limit, ask your admin for an increase" :') I ask them to optimize their crap instead...

[WKups @kups_w]

@thsottiaux Man, using codex these days is an absurd. I remember when I was able to run gpt 5.5 on medium for the whole week, facing 5h limit after 2h, codding all the time.

Now it is like I spent the whole 5x limit after 4h working on 5.6 SOL on MEDIUM. what the hell...

[WKups @kups_w]

@Gegam245074 how do you know it was opus 5.2, not just 5 :)