Gtol

[Gtol @gtolontop]

@theo fix auto compaction please

[Gtol @gtolontop]

@thsottiaux Worse thing gpt6 is gpt5.6 but cost less for u its just an improvment for ur server ??????? not for us bcs gpt6.1 sol is gpt6 astra minor with rebranding bruhhhhh noob

[Gtol @gtolontop]

@jullerino possible to fix auto compact pls

[Gtol @gtolontop]

haiku vs luna

[Gtol @gtolontop]

@maria_rcks @theo why autocompact not work ? Claude have auto compact set a 300k but it go up to 421k

[Sai @sai_borg]

🎉 Introducing the Sai API

Build your own Muse, assistant, or whatever you want, with one API call:

> A Windows cloud computer it can call its own
> Tasks in plain English, no clicking scripts
> Asks first before the scary stuff

Try it now → https://t.co/dVCe161bIu

[Gtol @gtolontop]

@DFintelligence @MistralAI après tes un zgeg aussi

[Ahhanya 🦩 @ik__6lemmon]

where do you see yourself in 1 hour?

[Gtol @gtolontop]

@theo its possible to take the draft from here and make a new conv? could be cool pls

[Gtol @gtolontop]

@DFintelligence Avec 46B actifs, 15T tokens et 0,5 PFLOP/s utiles/GPU, 6ND donne ~25 jours sur 3 800 GPU. Calcul simplifié, mais rien d’impossible a priori. Il manque surtout la durée et le nombre de tokens pour juger.

[Gtol @gtolontop]

@theo the app is insane thx

[Gtol @gtolontop]

Mistral is starting to host OSS models -> GLM 5.2

Here are the performance results of their model vs. GLM 5.2 it looks like a damn RL on GLM 5.2 with more resources; I'm blown away

[Gtol @gtolontop]

@MistralAI its so bad hahahahah im dead

[Gtol @gtolontop]

@leploutos débat de golmon

[Gtol @gtolontop]

Mistral release a good model wtf

[Gtol @gtolontop]

@Angaisb_ from where you find this ?

[Gtol @gtolontop]

@MistralAI benchmark ?

[Gtol @gtolontop]

@pmitu sexual intelligence

[Gtol @gtolontop]

@Angaisb_ @theo idk why ppl defend oai lol opus is just so better, sol6.1 is opus medium without taste etc lol

[Gtol @gtolontop]

@Angaisb_ oai is just dead rn lol