Be ready - October 14th

Bro, you’re literally a one-man army.

You made a faster finetune of Qwen 3.8 Flash Next without it being dumber than iq3_s?

🙃

You missed the chat bit ;)

I'm not ready

bro needs a pro video designer whats is this lmao

Indeed

@Gencaykazo Cool rendering, sharp and funny, loving it

@coldniko Yeah all good but the project deserves better if you kwim

👀

Niko, you are not an AGI come from future, right?

I wish, AGI doesnt have to sleep

Amazing work you're doing mate

qwen 4? strata for minimax?

qwen4？

so many false claimes today about AI , I didnt beleive STRATA until i tried it in old server z620 100gb ram and 2 of rtx 3060 12gb even one work and its fast 40 t/s thats crazy i like also the calibration ..thanks Niko its great job

cool !!

you fixed the ram cache on dual node cpus. you said you would, in a short sentence and you did. that's a very valuable thing. i hope you take care of yourself and your rare personality. thank you Niko 🫵👊🤲

... no drones fly into your data 🚀

Bit cheesy but LFG. We'll improve in the future.

🤣

👍👍👍

Caught the GUI! Hope it will be able to have web search ability and do deep research and other agent capabilities.

Already ready!

👀

Shoot I might cancel my Mac Ultra order. Are you working on Mac MLX version of Strata?

Boost Loading 🚀😀

Thank you very much Bro,I got 160 token/s on RTX 6000ada.

Niko.. quick question what should I buy an M5, spark or just a dual 3090 with 256gb ram

256gb ram

🙏

kaggle tpu 11s strata(rtx3080ti) 15.62s The results of running the same benchmark were identical in terms of time taken and performance.

Never been so ready.

👀

太棒了

WOWWWW!!! 🦾🚀

At this speed, are you going IPO on 10/14?

More models please, but I need more compute 🤡🥲 Running Qwen3.8-Flash-Next-IQ3_S at 20k ctx ~30 tok/s is pretty nice !

mac platform support？？ mac and nvda cards mixture？？

Nice :)

When IPO dude. congrats