@waefrebeorn i've been using the same model just the q4 xl quant, 72k ctx window with vision turned off and kv quan at fp8 on llama cpp, 8gb 4060 and 24 system ram, i get like 30tps write speed, genuinely love this model for agentic use or basic coding
Game developers are being attacked through hacked friends, they are sending fake game files and taking over your computer; be careful -
A fellow dev lost their account from attacks like this, and we're trying to get it back:
https://t.co/6uf3Pix9if
#IndieDevs#indiegamedev@TeamYouTube
#YouTubeHack
My Gmail ([email protected]) & YouTube channel (https://t.co/IT37UdLQb4) were phished/hacked from Istanbul, Turkey, passwords changed, parental controls set to unknown Turkish email. Google closed case despite US-only activity & evidence.
I followed up via email to [email protected] with new details from my State Police & FBI reports (plus details about other victims affected). Please escalate for re-evaluation. The hacker/phisher still in control risks more attacks.
@TeamYouTube #YouTubeHack #AccountRecovery
@grok@chandrarsrikant@grok Pull benchmark results you can find for 30B and 105B, and compared them to similar sized models like glm 4.7 flash for 30B and qwen3 next 80B for 105B
There are so many ways to enter to win FREE VIDEO CREATION GEAR and RESOURCES! ...and you can keep entering every single day until June 13th!
ENTER HERE: https://t.co/8GgjoEjIbL
Thanks to our sponsors @TetherTools@CINECOLOR_IO#5DayDealVideo#5daydifference#VideoCreator