AHAHAHAHAHHAHAHAH
That's AA2.1 idiot
Who the fuck cares.
still not using a closed-source model... sorry!
>>110371796the only thing that really matters is which local or chinese model can produce like Claude Opus 5.5.
>>110372219OP is a poorfag he can't run local
Wow, who gives a flying fuck.
>>110372239no i can't run claude opus 5.5 local and neither can anyone else you retard
>>110372219>the only thing that really matters is which local or chinese model can produce like Claude Opus 5.5.Probably the next GLM model.
>>110371796Conversation is the only true measure of LLM intelligence.
>>110372939Coding is most important because it allows you to build better AI and harness.
>>110371796How the fuck did Meta and Google get this good so quickly?Their models were a huge fucking joke a few months ago, even worse than some random chink shit at the time
>>110371796ah yes a coding race where everyone is actively stealing each others progress anyway
>>110373086>How the fuck did Meta and Google get this good so quickly?They're distilling Claude models. oAI is the only company that doesn't and they fell behind.
>>110372834Can GLM be uncensored?
>>110371796a while ago i cancelled my google ai subscription to try out open ai, after a month i cancelled it and went back. this turn of events does not surprise me. how dare you lecture me about which ai service is good, i have literally seen primagen vibecoding on stream. i know what level he's playing at. do not take that fucking tone with me, bud. last warning.
>>110373580yes, there is even a version for vramlets: https://huggingface.co/peasantsmith/GLM-5.3-Flash-Maya-M-Derisked-IQ2_S-GGUF
>>110371796I have a $20 plan both for Claude and Codex, it would be funny if I will ever need to swap Codex with Antigravity
>>110373638Thanks kind Anon