Panaiq

Building solutions with ai

[Panaiq @panaiq_ai]

@JustLingonberry 🤣 ipo delayed

[Panaiq @panaiq_ai]

@Biggest @ItsmeAjayKV @QwenDevs Is it like qwen4 - a3b

[Panaiq @panaiq_ai]

@Nozelcode I have my own video harness for it … https://t.co/NPKVrdfHTF

[Panaiq @panaiq_ai]

@MiaAI_lab Qwen 4 series will come soon .. my x feed is filled with with it

[Panaiq @panaiq_ai]

@agentspanel @typesafeai I see one named laya with same api structure as jev

[Panaiq @panaiq_ai]

@MiaAI_lab Specifically on artificial analysis..

[Panaiq @panaiq_ai]

@MiaAI_lab Then how these models gets in the top ranking in every arena..

[Panaiq @panaiq_ai]

@GoSailGlobal Thanks for sharing i build something similar but looks like you have a opensourced version of it

[Panaiq @panaiq_ai]

@SDDFounder Its a replica https://t.co/gEVzFmp3UF

[Panaiq @panaiq_ai]

@jauneclaude @MiaAI_lab @Italianclownz @ghazni1010 Sure let me try now

[Panaiq @panaiq_ai]

@jauneclaude @MiaAI_lab @Italianclownz @ghazni1010 here it is

[Panaiq @panaiq_ai]

@Italianclownz I just purchased a amd 9700 ai pro on windows llama cpp getting me around 30-40 token tg for agentic work at 100k context

[Panaiq @panaiq_ai]

@jauneclaude @MiaAI_lab @Italianclownz @ghazni1010 I did the same setup with llama cpp on q4_kL and works decent around 30 token/sec with 100k context for agentic work(very long prompts) just wondering if this the maximum i can get put of this model

[Panaiq @panaiq_ai]

@rege_dev Looks so good.. what ma the setup hyperframe or anything else

[Panaiq @panaiq_ai]

@ItsmeAjayKV @Alibaba_Qwen @UnslothAI For some reason i didn’t find dsh that good can u tell me what plugins u mostly work with

[Panaiq @panaiq_ai]

@twistartups @Jason @Lons @ThisWeeknAI @LAUNCH If jensen can give the electricity bill .. thats a plan

[Panaiq @panaiq_ai]

@timkochjar @rauchg @Anthropic @Remotion Which video you are talking about ! Is it like complete claude code motion design!!

[Panaiq @panaiq_ai]

@LordyVisuals Done followed send it

[Panaiq @panaiq_ai]

@ariskaa_ai Will it be a frontline model with 2-3T param or for commercial audiences.. like a 27B